developing ohmic contacts to gallium nitride · 2016-10-12 · accelerated life testing (alt) was...

Developing Ohmic Contacts to Gallium Nitride

for High Temperature Applications

by

Shirong Zhao

A Dissertation Presented in Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Approved July 2016 by the

Graduate Supervisory Committee:

Srabanti Chowdhury, Chair

Stephen Goodnick

Yuji Zhao

Robert Nemanich

ARIZONA STATE UNIVERSITY

August 2016

i

ABSTRACT

Gallium Nitride (GaN), being a wide-bandgap semiconductor, shows its

advantage over the conventional semiconductors like Silicon and Gallium Arsenide for

high temperature applications, especially in the temperature range from 300°C to 600°C.

Development of stable ohmic contacts to GaN with low contact resistivity has been

identified as a prerequisite to the success of GaN high temperature electronics. The focus

of this work was primarily derived from the requirement of an appropriate metal contacts

to work with GaN-based hybrid solar cell operating at high temperature.

Alloyed Ti/Al/Ni/Au contact and non-alloyed Al/Au contact were developed to

form low-resistivity contacts to n-GaN and their stability at high temperature were

studied. The alloyed Ti/Al/Ni/Au contact offered a specific contact resistivity (ρc) of

6×10-6 Ω·cm2 at room temperature measured the same as the temperature increased to

400°C. No significant change in ρc was observed after the contacts being subjected to

400°C, 450°C, 500°C, 550°C, and 600°C, respectively, for at least 4 hours in air. Since

several device technology prefer non-alloyed contacts Al/Au metal stack was applied to

form the contacts to n-type GaN. An initial ρc of 3×10-4 Ω·cm2, measured after

deposition, was observed to continuously reduce under thermal stress at 400°C, 450°C,

500°C, 550°C, and 600°C, respectively, finally stabilizing at 5×10-6 Ω·cm2. Both the

alloyed and non-alloyed metal contacts showed exceptional capability of stable operation

at temperature as high as 600°C in air with low resistivity ~10-6 Ω·cm2, with ρc lowering

for the non-alloyed contacts with high temperatures.

The p-GaN contacts showed remarkably superior ohmic behavior at elevated

temperatures. Both ρc and sheet resistance (Rsh) of p-GaN decreased by a factor of 10 as

ii

the ambient temperature increased from room temperature to 390°C. The annealed Ni/Au

contact showed ρc of 2×10-3 Ω·cm2 at room temperature, reduced to 1.6×10-4 Ω·cm2 at

390°C. No degradation was observed after the contacts being subjected to 450°C in air

for 48 hours. Indium Tin Oxide (ITO) contacts, which has been widely used as current

spreading layer in GaN-base optoelectronic devices, measured an initial ρc [the resistivity

of the ITO/p-GaN interface, since the metal/ITO ρc is negligible] of 1×10-2 Ω·cm2 at

room temperature. No degradation was observed after the contact being subjected to

450°C in air for 8 hours.

Accelerated life testing (ALT) was performed to further evaluate the contacts

stability at high temperatures quantitatively. The ALT results showed that the annealed

Ni/Au to p-GaN contacts is more stable in nitrogen ambient, with a lifetime of 2,628

hours at 450°C which is approximately 12 times longer than that at 450°C in air.

iii

ACKNOWLEDGMENTS

First of all, I would like to express my deepest appreciation to my thesis

supervisor, Prof. Srabanti Chowdhury, for introducing me to this interesting field and for

her advice, inspiration, encouragement, and continuous support throughout my PhD. Her

enthusiasm on the research and her mission for providing high-quality work have made a

deep impression on me. I have learnt not only the knowledge in research, but also the

dedicating spirit for the work. I am really glad to be associated with a person like Prof.

Chowdhury in my life.

I would also like to thank my committee members, Prof. Stephen Goodnick, Prof.

Yuji Zhao, and Prof. Robert Nemanich, for giving professional guidance, valuable

discussion, and careful corrections. This thesis benefits a lot from their support.

Special thanks must go to Prof. Fernando Ponce and his TEM group for providing

TEM analysis which is an important part of my thesis. I have learnt a lot from our

discussions and I really enjoy collaborating with them.

I am indebted to all the members in Prof. Chowdhury’s research group that I have

interacted with over the years. I have learned so much both from their work and from our

everyday conversations. Their timely help and friendship shall always be remembered.

I am grateful to the support from the Center for Solid State Electronics Research

at ASU, where most experiments were carried out, and the LeRoy Eyring Center for

Solid State Science at ASU for providing the characterization service.

I gratefully acknowledge the funding support from the Advanced Research

Projects Agency-Energy (ARPA-E).

iv

TABLE OF CONTENTS

Page

LIST OF TABLES ............................................................................................................. vi

LIST OF FIGURES .......................................................................................................... vii

LIST OF SYMBOLS ....................................................................................................... xiii

CHAPTER

1 INTRODUCTION OF GAN HIGH TEMPERATURE ELECTRONICS ....... 1

1.1 Demands for High Temperature (HT) Electronics .................................. 1

1.2 Advantages of GaN for HT Electronics .................................................. 3

1.3 High Temperature InGaN Topping Cells for Hybrid Solar Converters .. 5

1.4 High Temperature Contacts to GaN ........................................................ 7

2 THEORY AND BACKGROUND OF OHMIC CONTACTS TO GAN ......... 9

2.1 A Primer for Semiconductor-Metal Contacts .......................................... 9

2.2 Schottky Barriers Heights of Metal Contacts to GaN ........................... 11

2.3 Summary of Ohmic Contacts to GaN at Room Temperature ............... 12

2.3.1 Ohmic Contacts to n-GaN ........................................................... 12

2.3.2 Ohmic Contacts to p-GaN ........................................................... 14

2.4 Challenge for High Temperature Stable Ohmic Contacts ..................... 18

3 EXPERIMENTAL STUDY AND METHODOLOGY TO DEVELOP

STABLE HT CONTACTS TO GAN ............................................................. 20

3.1 Transmission Line Method for Characterizing Contact Resistivity ...... 20

3.2 Fabrication of Contacts to GaN ............................................................. 23

3.3 Contacts Characterization ..................................................................... 25

v

CHAPTER Page

3.4 Summary ............................................................................................... 27

4 DEVELOPING OHMIC CONTACTS TO N-GAN FOR HT

APPLICATIONS ............................................................................................ 29

4.1 Alloyed Ti/Al/Ni/Au Contacts .............................................................. 29

4.2 Non-Alloyed Al/Au Contacts ................................................................ 38

4.3 Summary ............................................................................................... 43

5 DEVELOPING OHMIC CONTACTS TO P-GAN FOR HT

APPLICATIONS ............................................................................................ 45

5.1 Annealed Ni/Au Contacts to p-GaN ...................................................... 45

5.1.1 Electrical Characterization .......................................................... 45

5.1.2 Microstructure Characterization.................................................. 57

5.2 Indium Tin Oxide Contacts to p-GaN ................................................... 62

5.3 Summary ............................................................................................... 70

6 LONG TERM RELIABILITY AT HIGH TEMPERATURE ........................ 71

6.1 Arrhenius Model for Accelerated Life Testing ..................................... 71

6.2 Accelerated Life Testing of the Annealed Ni/Au to p-GaN Contact .... 73

6.3 Summary ............................................................................................... 78

7 CONCLUSION AND FUTURE WORK ....................................................... 79

7.1 Conclusion ............................................................................................. 79

7.2 Future Work .......................................................................................... 80

REFERENCES ................................................................................................................. 83

vi

LIST OF TABLES

Table Page

1. Values of ρc for Contact Sample A, B, and C as a Function of Time under

Thermal Stress at 475 °C, 500 °C, and 525 °C, respectively, in Air..........….…..74

2. Comparison of Contact Lifetime (Hours) in Air and in N2 at Temperatures from

300 °C to 500 °C ……...…………………………………………………………78

vii

LIST OF FIGURES

Figure Page

1. High-Temperature Applications with Ambient Temperature Range from 300 °C

to 600 °C. ................................................................................................................ 2

2. Semiconductor Intrinsic Carrier Concentration as a Function of Temperature. ..... 4

3. Estimated Practical Electrical Efficiency Limit of a PV Topping Hybrid Solar

Converter at 100X Concentration, versus PV Temperature. Dashed Curves

Compare the Dispatchable Electricity from Heat without PV and in the Hybrid

System. Inset Schematic Shows the Hybrid Converter Configuration ................... 6

4. Metal and n-type Semiconductor Pair before Contact (a) and after Contact (b).

The Metal Work Function ϕm is Greater than That for the Semiconductor, ϕs. .... 10

5. Metal and p-type Semiconductor Pair before Contact (a) and after Contact (b).

The Metal Work Function ϕm is Smaller than That for the Semiconductor, ϕs. ... 11

6. High-Resolution Transmission Electron Microscopy Image of the Interfacial Area

of the Ti/Al/Ni/Au Contacts to n-GaN after Annealing........................................ 13

7. High-Resolution TEM Image of Ni/Au Contact after Annealing in Air Ambient

for 20 min at Temperature of 500°C. .................................................................... 17

8. Schematic Drawings of Microstructure at p-GaN/Metal Interface of Ni/Au

Contacts, As-Deposited, Annealed in N2 and Annealed in Air at Temperature of

500~600 °C. .......................................................................................................... 18

9. Schematic Diagrams of Metal Contacts Patterned with TLM Structure on GaN

Surface and Measurement of I-V Curves with Four-Point Probes. ...................... 21

10. Plot of Total Resistance as a Function of TLM Pad Spacing. .............................. 21

viii

Figure Page

11. The Current Flow through the Contacts and the Semiconductor. ......................... 23

12. The Layer Structure and Doping Profile of the n-GaN and p-GaN Samples Used

for HT Contact Studies. ........................................................................................ 24

13. Process Flow of the Contact Patterned with TLM Structures............................... 25

14. Four-Probe I-V Characterization System with a Thermal Chuck Which Can Heat

the Sample up to 400 °C during the Measurement in Air. .................................... 26

15. A Mini-Brute Furnace for HT Stress with N2 and Air Flow................................. 27

16. Electrical Characteristics of the Alloyed Ti/Al/Ni/Au Contacts to n-GaN

Measured at Room Temperature in Air, (a) I-V Curves Measured between

Adjacent TLM Pads with Spacing Increasing from 3um to 30um, and (b) Plot of

Measured Resistance versus TLM Contact Spacing. ............................................ 29

17. Temperature Dependent I-V Characteristics of the Alloyed Ti/Al/Ni/Au Contacts,

Taken Between Two Adjacent Pads with Gap Spacing 10um. ............................ 31

18. Contact Resistivity and Sheet Resistance of the Alloyed Ti/Al/Ni/Au to n-GaN

Contacts as a Function of Measurement Temperature. ......................................... 31

19. Contact Resistivity and Sheet Resistance of the Ti/Al/Ni/Au to n-GaN Contacts as

a Function of Time under 400 °C. ........................................................................ 32

20. Microscopy Images of the Alloyed Ti/Al/Ni/Au Contacts Surface, Pre (a) and

Post (b) the 450 °C Bake, respectively. Plots of Measured Resistance versus TLM

Contact Pads Spacing, Pre (c) and Post (d) the 450 °C Bake, respectively. ......... 34

ix

Figure Page

21. The Change of Contact Resistivity and Sheet Resistance of the Alloyed

Ti/Al/Ni/Au Contacts after Thermal Stress at 400 °C, 500 °C, 550 °C, and 600 °C

for 4 Hours Each with Ambient Air. ..................................................................... 35

22. The Change of Contact Resistivity and Sheet Resistance of the Alloyed

Ti/Al/Mo/Au Contacts after Thermal Stress at 400 °C, 500 °C, 550 °C, and

600 °C for 4 Hours Each with Ambient Air. ........................................................ 36

23. Cross-Section TEM Images of the Alloyed Ti/Al/Ni/Au Contacts to n-GaN. (a)

Low Magnification Image of the Interface and (b) High Resolution Image of the

Interfacial TiN Layer. ........................................................................................... 37

24. Electrical Characteristics of the As-Deposited Al/Au Contacts to n-GaN

Measured at Room Temperature in Air, (a) I-V Curves Measured between

Adjacent TLM Pads with Spacing Increasing from 3um to 30um, and (b) Plot of


25. The Change of Contact Resistivity and Sheet Resistance of the Al/Au Contacts

after Thermal Stress at 400 °C, 500 °C, 550 °C, and 600 °C for 4 Hours Each with

Ambient Air. ......................................................................................................... 40

26. Cross-Section TEM Image of the Al/Au Contacts to n-GaN after Thermal Stress

at 600 °C in Air for 4 Hours. ................................................................................ 41

27. STEM Image of the Interfacial Area on the Al/Au to n-GaN Contacts after

Thermal Stress at 600 °C in Air for 4 Hours. ....................................................... 42

x

Figure Page

28. (a) EELS Scan Trace, across the 66 nm Interfacial Layer and (b) EELS Spectrum

of N, O, Ga, Al, and Au on the Al/Au to n-GaN Sample after Thermal Stress at

600 °C in Air for 4 Hours. .................................................................................... 42

29. Extracted Contact Resistivity as a Function of Rapid Thermal Annealing Time at

500 °C, 550 °C, and 600 °C, respectively. ............................................................ 46

30. Measured Resistance versus TLM Pad Spacing after the Ni/Au to p-GaN Contacts

were Annealed at 600 °C in N2 for 3 min. The Inset Picture Shows the I-V Curves

Measured between Adjacent Contact Pads with Various Spacing. ...................... 46

31. Temperature Dependent I-V Characteristics of the Annealed Ni/Au to p-GaN

Contacts, Taken between Two Adjacent Pads with Gap Spacing 5um. ............... 48

32. Temperature Dependent Specific Contact Resistivity ρc, (a), and Sheet Resistance

Rsh, (b), of the Annealed Ni/Au to p-GaN Contacts. ............................................ 49

33. Measured Resistance versus TLM Pad Spacing after the Annealed Ni/Au Contacts

Samples were Subjected to 450 °C in Air for 48 h. The Inset Picture Shows the I-

V Curves Measured between Contact Pads with Various Spacing from 3um to

30um. .................................................................................................................... 51

34. Time Evolution of the Specific Contact Resistivity ρc, (a), and Sheet Resistance

Rsh, (b), of the Annealed Ni/Au Contacts over the 48-hour Thermal Stability Test

at 450 °C in Air. .................................................................................................... 53

xi

Figure Page

35. (a) I-V Characteristics Measured at Room Temperature and Taken between Two

Adjacent TLM Pads with Gap Spacing 5 um and (b) the Extracted ρc of the Ni/Au

Contacts after Being Subjected to Thermal Stress at 500 °C, 550 °C, 600 °C, and

650 °C in Series in Air. ......................................................................................... 55

36. The Extracted Rsh of the Ni/Au Contacts after Being Subjected to Thermal Stress

at 500 °C, 550 °C, 600 °C, and 650 °C in Series in Air. ...................................... 56

37. I-V Characteristics of the Ni/Au Contacts Measured at Room Temperature after

Being Subjected to 700 °C for 4 hours in Air. ...................................................... 56

38. Cross-Section TEM Image of the Ni/Au Contacts after Post-Deposition Annealing,

Showing Three Layers with Two Interfaces. ........................................................ 58

39. High-Resolution TEM Image of the Top Interface, Marked as Interface 1, of the

Annealed Ni/Au Contacts. .................................................................................... 59

40. High-Resolution TEM Image of the Metal-GaN Interface, Marked as interface 2,

of the Annealed Ni/Au Contacts. .......................................................................... 59

41. Cross-Section TEM Image of the Ni/Au Contacts after Thermal Stress at 700 °C

for 4 hours in Air................................................................................................... 61

42. High-Resolution TEM Image of the Interfacial Area on the Ni/Au Contacts after

Thermal Stress at 700 °C for 4 hours in Air. ........................................................ 61

43. Two TLM Structures to Test the Contact Resistivity of ITO to p-GaN, (a), and the

Sheet Resistance of ITO Film, (b), Respectively. ................................................. 63

xii

Figure Page

44. (a) I-V Characteristics of the ITO to p-GaN Contact after Post-Deposition

Annealing at 500 °C in Air for 1h, Measured on TLM A, and (b) Plot of


45. (a) I-V Characteristics of the ITO to p-GaN Contact after Thermal Stress at

450 °C in N2 for 8h, Measured on TLM A, and (b) Plot of Measured Resistance

versus TLM Contact Spacing................................................................................ 66


450 °C in Air for 8h, Measured on TLM A, and (b) Plot of Measured Resistance



500 °C in Air for 8h, Measured on TLM A, and (b) Plot of Measured Resistance


48. Increase of ρc as a Function of Time under Thermal Stress at 475 °C, 500 °C, and

525 °C, Respectively, in Air. ................................................................................ 74

49. Arrhenius Plot of ln (tf) vs 1/kT in Air. ................................................................. 75

50. Estimated Lifetime in Air versus Temperature according to Arrhenius Model. .. 76

51. Arrhenius Plot of ln (tf) vs 1/kT in N2. .................................................................. 77

52. Estimated Lifetime in N2 versus Temperature according to Arrhenius Model. ... 77

xiii

LIST OF SYMBOLS

Symbol Page

1. ρc, Specific Contact Resistivity ……………….…………………………………13

2. Rsh, Sheet Resistance …………………………..………………………………...22

3. Ea, Activation Energy ………………………………………………………….71

4. tf, Lifetime ………………………………………...……………………………..72

1

Chapter 1

INTRODUCTION OF GAN HIGH TEMPERATURE ELECTRONICS

1.1 Demands for High Temperature (HT) Electronics

It is increasingly recognized that semiconductor based electronics that can

function at ambient temperatures higher than 150 °C without external cooling could

greatly benefit many important applications, especially in the automobiles, aircrafts, gas

turbines, oil exploration, and space exploration industries[1]–[5]. More than 40,000 oil

and gas wells are being logged every year by the oil industry where electronic and

electrical parts range from cabling, connectors, and passive devices to A/D converters,

operational amplifiers, motors and solenoids are needed. The temperature in the oil and

gas wells can range up to 300 °C for long periods of time during logging. Operational

temperatures for the geothermal wells are generally higher than those for oil well and can

range up to 400 °C for wells being drilled to depths greater than 11 km. The temperature

needs of the automotive industry can range from 150 °C to 800 °C. The lower

temperature range, 150 °C to 300 °C, for electronics would include engine block

embedded sensors to monitor cylinder pressure and temperature, sensors in brake pads,

transmissions, and in chemical fluid reservoirs, while the exhaust gas sensor represents an

extreme temperature requirement, 800 °C. The aircraft industry has recognized since the

mid-1960s, that placing monitor and control electronics directly on the jet engines could

provide many benefits. Electronics capable of operating at 300 °C would satisfy most

engine monitoring and control needs, while the engine temperature of high performance

aircraft, e.g. supersonic flight at Mach 5, could range from 550 °C to 650 °C. Scientific

space missions to other planets and deep space also need high temperature (HT)

2

electronics. The 465 °C electronic devices are needed for instrumenting Venus surface

probe mission, where the maximum surface temperature can reach 465 °C. Fig. 1 shows

the major high-temperature applications with ambient temperature range from 300 °C to

600 °C.

Fig. 1 High-temperature applications with ambient temperature range from 300 °C to 600

°C.

High temperature harsh environments are extremely difficult for electronics to

function properly in-situ. Traditionally, engineers rely on active or passive cooling when

designing electronics that must function above typical high (150 °C) temperature ranges,

or remotely locating the electronics from the high-temperature region. However, active

cooling is not always a desired option in many applications or the cost can be very high.

The ability to operate electronic systems at high temperatures will not only make new

products possible but it will also decrease the cost and increase the reliability of current

3

products by removing the need for large, massive, complex cooling systems and the

cabling and interconnections required for remote placement of the electronics.

1.2 Advantages of GaN for HT Electronics

The temperature limit of semiconductor materials depends highly on their

bandgaps. The conventional narrow bandgap silicon devices cannot be operated reliably

above 150 °C due to the generation of thermal carriers which can cause leakage, thermal

runaway and latch-up at reverse bias[1]. The control of the carrier concentration is crucial

for the performance of semiconductor devices. The intrinsic carrier concentration for

several semiconductors as a function of temperature is shown in Fig. 2[6], [7]. At

temperatures above 300 °C, wide bandgap semiconductor (SiC, GaN, diamond, and AlN)

with Eg over 3 eV have much lower intrinsic carrier concentrations compared to Si and

GaAs and, thus, do not run into intrinsic carrier conductivity related issues until much

higher temperatures. This implies that devices designed for operating above 300 °C

should be fabricated from wide-bandgap semiconductors to avoid the deterioration effects

of thermally generated carriers.

4

Fig. 2 Semiconductor intrinsic carrier concentration as a function of temperature[6].

SiC-based and GaN based devices are currently the most matured wide-bandgap

semiconductors. SiC and GaN technologies are attractive for harsh environment

operation because of their high chemical and physical stability; higher atomic

displacement energy than many other semiconductor materials. It was reported that the

practical temperature limit of GaN and SiC are 600 °C[2], sufficient to support long-term

device operation at high temperature. Mass-produced single-crystal SiC wafers have been

commercially available while GaN crystals have mostly been grown heteroepitaxially on

foreign substrates like sapphire and SiC. SiC crystals have much fewer defects than GaN.

Thus, it is not surprising that more research aimed at HT devices has been carried on SiC

than GaN. However, GaN also has some important advantages over SiC, i.e.

5

heterojunction device design capability and a direct energy band gap. The AlGaN/GaN

interface forms easily a high-electron mobility two-dimensional electron gas (2DEG) and

offers very low channel resistance while SiC devices show rather high channel resistance.

GaN-base LED has much higher emission efficiency than SiC because GaN has a direct

bandgap whereas SiC has an indirect bandgap which is unfavorable for light emission. In

addition, GaN (Eg=3.4 eV) and its alloy with InN (Eg=0.7 eV) has a band gap that is

continuously tunable from 0.7 to 3.4 eV[8], which covers a broad solar spectrum from

near-infrared to near-ultraviolet wavelength region, offering a great advantage in design

and fabrication of high efficiency devices for photovoltaic applications[9], [10]. It is well

known that GaN is dominating the lighting and RF industries and entering the power

electronics market. Therefore, HT GaN electronics will bring unprecedented advantages

in all these applications.

1.3 High Temperature InGaN Topping Cells for Hybrid Solar Converters

Photovoltaic (PV) and concentrating solar power (CSP) are the two main ways of

harvest solar energy. However, PV is unlikely to economically supply much more than

10% of the world’s electricity due to the high cost of electricity storage[11]. In contrast,

CSP collects heat, store heat, and dispatch electricity day and night, but it requires very

large and expensive power plants. The hybrid solar converters which combines PV and

CSP can address the issue of solar energy storage, and also optimally exploit the solar

spectrum to realize higher conversion efficiencies and low electricity costs, while

ensuring the availability of inexpensive dispatchable solar power[11]–[13]. The estimated

practical electrical efficiency limit of a PV topping hybrid solar converter at 100X

concentration, versus PV temperature is shown in Fig. 3[11]. The total efficiency of the

6

hybrid solar converter is higher than both PV and CSP alone. Despite the drop of PV

efficiency at high temperature, the efficiency of CSP increases faster with temperature,

thus the total efficiency increases with temperature. Also, the dispatchable fraction of

electricity is above 50% when the hybrid solar converter is operating above 400 °C.

Fig. 3 Estimated practical electrical efficiency limit (black, solid) of a PV topping hybrid

solar converter at 100X concentration, versus PV temperature. Dashed curves compare

the dispatchable electricity from heat without PV (green) and in the hybrid system

(black). Inset schematic shows the hybrid converter configuration[11].

One challenge of the hybrid solar converter is that PV cells, based on wide-

bandgap materials, need to be durable for long term operation at high temperatures. GaN

and its ternary nitride alloy InxGa1-xN has the potential application for the HT PV cells

due to the wide-bandgap and tunable bandgaps in the range of 0.7 eV and 3.4 eV,

covering most of the range of solar spectrum. While GaN alone can withstand high

7

temperatures above 600 °C, a major challenge is to develop reliable low resistivity ohmic

contacts to the device itself. At temperatures higher than 300 °C, failure of the entire cell

can occur due to the diffusion of the contact metals, oxidation and reactions at the metal-

semiconductor interface. Therefore, the development of thermally and electrically stable

ohmic contacts for HT operation is critical for HT GaN electronics, including the hybrid

InGaN solar cells.

1.4 High Temperature Contacts to GaN

By contrast with the Si and GaAs devices, which operating temperature is limited

by the electronic properties of the semiconductor material, the maximum operating

temperature of GaN devices is limited by stability of the contacts. Some device

parameters such as response time, output power and etc. depend strongly on the ohmic

contact resistivity and its stability at high operating temperatures. For example, in the

InGaN PV devices, contact resistance is part of the series resistance. Increase in the series

resistance would reduce the fill factor of the PV cells, and thus reduce the efficiency of

the InGaN PV cells[14]. Therefore stable and low resistivity ohmic contact is considered

as the critical factor determining GaN-based device performance at high temperature. The

thermal properties and long term ageing of the ohmic contacts to SiC have been well

investigated[15]–[19], which provides insights into approaches to evaluation high-

temperature metal contacts to GaN. An InGaN solar cell operating at temperatures in

excess of 300 °C, requires thermally stable contacts to both n-type GaN and p-type GaN

capping layers. The focus of this work was specifically derived from the need for

appropriate ohmic contacts to both n-type GaN and p-type GaN at elevated temperatures

up to 450 °C, also trying to push the contacts temperature limit to 600 °C. The

8

temperature-dependent contact resistivity of each contacts, their long term stability under

high temperature aging, and possible contact degradation mechanism are discussed. We

also introduced a method to evaluate the lifetime of ohmic contacts at elevated

temperatures using accelerated life testing.

The success of this work enables the possibility of high temperature operation of

InGaN solar cell up to 450 °C and unipolar GaN devices, e.g. AlGaN/GaN HEMTs, up to

600 °C in air ambient, making GaN HT devices good supplements to SiC HT electronics.

We will be able to use the dispatchable and economic solar energy with better efficiency-

to-cost ratio. The work discussed on contacts will significantly benefit HT power

electronics and RF electronics where thermal management and thermal stability are of

high importance.

9

Chapter 2

THEORY AND BACKGROUND OF OHMIC CONTACTS TO GAN

2.1 A Primer for Semiconductor-Metal Contacts

Semiconductor device must be connected through metal contacts to the outside

world with no adverse change to its current-voltage characteristics and minimum voltage

drop across the contact. This can be accomplished through low-resistance ohmic contacts

to the semiconductor. Ohmic contacts in a semiconductor are formed such that the

majority carriers can flow into the device from an external source (like a voltage or a

current supply) with minimal resistance. Conventionally metals with appropriate work

functions are chosen to accomplish no barrier to the majority carrier. However in

wideband gap semiconductors such as GaN due to lack of metals with appropriate work

function or surface pinning effect of the Fermi level, a Schottky based ohmic contact is

often more common, where majority carrier transport occurs through the barrier by

tunneling. Such designs involve heavy doping in the semiconductor surface so that the

carriers can tunnel through the barrier.

The band alignment effect for n-type semiconductor and p-type semiconductor are

illustrated in Fig. 4 and Fig. 5, respectively. For metal contact to the n-type

semiconductor, the barrier height ϕB (before the image force lowering effect) to electrons

traveling from metal to semiconductor is q(ϕm- χ) while that to electrons traveling from

semiconductor to metal is q(ϕm- ϕs), ϕm the metal work function, ϕs the semiconductor

work function, and χ the electron affinity. Similarly, for metal contact to the p-type

semiconductor, the barrier height to holes traveling from metal to semiconductor is

Eg+q(χ-ϕm), while that to holes traveling from semiconductor to metal is q(ϕs-ϕm).

10

Therefore, metal with lower work function is preferred to achieve contact to n-type

semiconductor, and metal with higher work function is preferred to achieve contact to p-

type semiconductor.

Fig. 4 Metal and n-type semiconductor pair before contact (a) and after contact (b). The

metal work function ϕm is greater than that for the semiconductor, ϕs.

11

Fig. 5 Metal and p-type semiconductor pair before contact (a) and after contact (b). The

metal work function ϕm is smaller than that for the semiconductor, ϕs.

2.2 Schottky Barrier heights of Metal Contacts to GaN and role of surface states

The above discussion indicates that the Schottky barrier height of metal-

semiconductor contact depends highly on the metal work function. However, many

researcher have found that the Schottky barriers of metal contacts to GaN depends

weakly on the metal work function[20]–[24].

The Schottky barrier heights of Ti, Au, Pd, Ni, Pt, with work functions of 4.3eV,

5.1eV, 5.1eV, 5.15 eV, 5.65 eV, respectively, on n-type GaN are measured to be

0.58eV[22], 0.94eV[24], 0.94eV[24], 0.99eV[23], and 1.04eV[24], respectively. These

experimental data illustrates that metal with lower work function has lower Schottky

barrier height to n-GaN. However, the difference in the Schottky barrier height among

these metals is significantly smaller than the difference in their work functions, implying

12

that the metal work function is not the only factor affecting the Schottky barriers of metal

contacts to n-GaN.

Similarly, the Schottky barrier heights of Ti, Au, Ni, Pt, with work functions of on

p-type GaN are measured to be 0.65eV, 0.57eV, 0.5eV, and 0.5eV, respectively[25].

Higher work function metal shows lower Schottky barrier height to p-GaN, but the

difference in Schottky barrier heights is small compared to that of metal work function.

The Schottky barrier of metal contacts to GaN is dominated by the fermi level

pinning at the surface due to GaN surface states. The minimum barrier height is around

0.5 eV for both n-GaN contact and p-GaN contact. Therefore, selection of appropriate

metal alone cannot form low resistivity contact. The presence of high concentration

carriers near the metal-GaN interface is essential to form ohmic contact to GaN.

2.3 Summary of Ohmic Contacts to GaN at Room Temperature

2.3.1 Ohmic Contacts to n-GaN

Si has been the preferred n-type dopant for GaN. It acts as a shallow donor with

activation energy ~15 meV. Heavy doping can be easily achieved with electron

concentration ~1020cm-3. As a result, ohmic contacts can be formed relatively easily by

using metals having a low work function. Ti and Al are commonly used for

semiconductor fabrications and they are also popular candidates for ohmic contacts to n-

GaN due to their low work functions, 4.33 eV and 4.28 eV, respectively. Ti and Al based

metallization schemes, such as Ti-only[26], Al-only[27], Ti/Al bilayer[27], [28],

Ti/Al/Ti/Au multilayer[29], and so on, have been used to form ohmic contacts to n type

GaN. It was widely believed that a solid phase reaction occurs between Ti and GaN

13

during the high temperature annealing[29]. Nitrogen out-diffuses from the GaN lattice to

form TiN and residual nitrogen vacancies act as donors in GaN. A high-resolution

transmission electron microscopy image of the interfacial area of the Ti/Al/Ni/Au

contacts to n-GaN after annealing is depicted in Fig. 6, showing the formation of a thin

poly-crystalline cubic TiN layer at the metal composite-GaN interface[30]. The

interfacial area therefore becomes heavily doped providing the configuration needed for

tunneling contacts[29]. Those contacts generally yields specific contact resistivity ρc

below 1×10-5 Ω cm2.

Fig. 6 High-resolution transmission electron microscopy image of the interfacial area of

the Ti/Al/Ni/Au contacts to n-GaN after annealing[30].

However, the Ti-only, Al-only, and the Ti/Al bilayer contacts to n-GaN are not

reliable for high-power and high-temperature device applications. Both Ti and Al tends to

be oxidized, which is unavoidable for HT electronics, especially for Ti. Besides, Al tends

14

to react with GaN and form wide-band-gap AlN at the interface after annealed at

600 °C[31]. Low melting points of Al (600 °C) presents another problem for thermal

stability of the Al-only and Ti/Al contacts. Therefore, those contacts will degrade after

high temperature annealing or operation, and thus can’t be used for HT electronics.

More complex metallization schemes, such as Ti/Al/Ni/Au multilayer[32]–[34],

Ti/Al/Pt/Au[35], [36], Ti/Al/Pd/Au[37], and Ti/Al/Mo/Au[38], are widely studied and

tested at high temperature for forming ohmic contacts to n-GaN and AlGaN/GaN HEMTs.

In these contact schemes, the low-resistivity Au coating metal is employed to improve the

resistance to oxidation of the Ti/Al layer during high temperature annealing. On the other

hand, the barrier metal (Ni, Pt, Pd, Mo) between Au and the Ti/Al layer is introduced to

prevent the inter-diffusion of Ti, Al, and Au[35], [39], and therefore improve the thermal

stability of the contact[40]. Hou et al. reported that alloyed Ti/Al/Pt/Au to n-type GaN

contacts was very stable at 600 °C over 10 hours in an air ambient [41]. The temperature

dependence of the specific contact resistivity, between 25 °C and 175 °C in Ti/Al/Ni/Au

ohmic contacts to n-GaN has been reported[42]. The specific contact resistance ρc

decreases with increasing measuring temperatures which is explained as the current

transport in the contacts occurs by a thermionic field emission mechanism.

2.3.2 Ohmic Contacts to p-GaN

In contrast to n-GaN, stable ohmic contacts with low resistivity to p-GaN are

much more difficult to achieve due to lack of suitable work function in metals, surface

states in GaN and deep acceptor [Mg] energy level. The strategy to make ohmic contacts

to p-type GaN is, in principle, to use metals having a large enough work function to

facilitate tunneling transport. However, p-type GaN has fundamental problems, which

15

make it difficult to form device-quality ohmic contacts with a specific contact resistance

that is lower than 1×10−4 Ω cm2. The first problem is the absence of appropriate metals

which have a work function that is large than that of the p-type GaN, with bandgap 3.4

eV and electron affinity 4.1 eV. Ni, Au, Pd, and Pt are widely used for p-GaN contacts,

and there work function are 5.15 eV, 5.1 eV, 5.12 eV and 5.65 eV, respectively. Thus, a

high Schottky-barrier is unavoidable on the metal-semiconductor interface. The second

problem arises from the difficulty in growing highly doped p-GaN, with hole

concentration above 1×1018 cm-3, due to the high activation energy, ~170 meV, of Mg. In

growth techniques such as MOCVD reactor where Hydrogen is present during growth,

formation of Mg-H complexes causes passivation of Mg[43]. However the passivation of

Mg can be recovered by annealing out the H from the crystal at 700 °C and therefore is

not a primary issue for high contact resistivity.

Metal schemes with large work function and also easy to use in device processing,

such as Ni, Au, Pd, and Pt, have been investigated in order to form low resistance ohmic

contacts to p-GaN. Ni/Au double layer contact technology to form ohmic contacts to p-

type GaN at room temperature has been widely reported, with a standard ρc ~1×10−3 Ω

cm2 after rapid thermal annealing (RTA) [44]–[56]. However, the mechanism of forming

ohmic contact after RTA is still not clear. Ho et al. reported that the decrease of the

contact resistivity was due to the formation of NiO during the annealing process in

air[45], and the NiO was found to be p-type conducting with hole concentration of 2×1017

Ω cm2[47]. Koide et al. proposed that the ohmic contact was achieved due to the increase

of the hole concentration in GaN surface layer caused by removal of hydrogen atoms

bonded with Mg acceptors during annealing[57], [58]. Fig. 7 shows a high resolution

16

image of the Ni/Au contact annealed in the air ambient at 500°C for 20 min. The Au layer

is directly in contact with the GaN surface with sharp interface. There is no NiO or other

intermixed layer formed at the interface. Ni diffuses away from the metal-GaN interface

and reacts with oxygen, forming a NiO compound on top of the Au layer. The Ni and Au

layer sequence is inverted after annealing and no reaction between the metals and the

GaN is observed at the interface.

Fig. 8 shows the schematic illustrations of the microstructural change of the

Ni/Au contacts after annealing in the N2 or air ambient. Before the metal deposition, there

is a thin contaminated layer with a thickness of around 2 nm, which is unavoidable as the

GaN epi-layer is exposed to air[59]. Therefore, the as-deposited metal is sitting on top of

the contaminated layer, likely Ga2O3 (n-type to semi-insulating in nature) offers a higher

contact resistivity. After annealing in the N2 ambient, Ni reacted with the contaminated

(oxide) layer and Au, removing the thin contaminated interfacial layer and forming an

Au–Ni solid solution which intimately contacted the p-GaN surface. Therefore, reduction

of the ρc value after annealing in the N2 ambient was believed to be due to the intimate

contact of the metal to the p-GaN surface. Annealing in the air ambient facilitated out-

diffusion of Ni to the Au surface, and Ni was then oxidized, forming the NiO on top of

the Au layer. Since the Au was not oxidized, the intimate contact of Au to p-GaN

occurred as a result of the out-diffusion of Ni to the surface.

17

Fig. 7 High-resolution TEM image of Ni/Au contact after annealing in air ambient for 20

min at temperature of 500°C[57].

18

Fig. 8 Schematic drawings of microstructure at p-GaN/metal interface of Ni/Au contacts,

as deposited, annealed in N2 and annealed in air at temperature of 500~600 °C[57].

2.4 Challenges in HT Stable Ohmic Contacts

The ohmic contacts are a critical factor that could restrict the HT device

application. The high operating temperatures may cause diffusion processes in the contact

layer and reactions between the contact components, which could lead to changes of the

contact properties during operation at high temperatures, and deterioration of the devices.

Also, possible reaction and diffusion between the metal and GaN occur at high

temperature which may result in reduction of carrier concentration on the surface of GaN

and thus degrade the ohmic contact. If the contact resistivity is not sufficiently low

19

inadmissible high voltage drop could arise due to the high current density in the contact

of high power devices. Hence, the following requirements to the ohmic contacts are

critical for GaN HT applications.

i) Low contact resistivity – in general, the make of low resistivity ohmic

contacts is difficult for wide band-gap semiconductors due to the difficulty

in doping and, in the case of p-type materials, due to the wide forbidden

band-gap.

ii) High temperature stability – this problem is very important in the wide

band-gap semiconductors. In the Si and GaAs devices the maximal

working temperature is limited by the material stability, because of that

the problem of the contact stability is important but not critical. The great

potential of III-V nitrides to work at temperatures up to 600 °C and higher,

set strong requirements for the thermal stability and reliability of the

contacts.

iii) Reproducibility – this requirement is important in the case of the device

production. Therefore the contact technology should allow the

achievement not only good performance, but good reproducibility.

The listed requirements point that the operation of high temperature and high

power GaN-based devices under severe conditions demands development of electrically,

thermally and chemically stable metal contacts.

20

Chapter 3

EXPERIMENTAL STUDY AND METHODOLOGY TO DEVELOP STABLE HT

CONATCTS TO GAN

3.1 Transmission Line Method for Characterizing Contact Resistivity

Transmission Line Method (TLM) was used for evaluating the ohmic contacts to

n and p type GaN throughout in this study. This technique was proposed by Reeves and

Harrison. Details about this method are available in their published work[60].

A schematic diagram of a semiconductor material with ohmic contact pads

prepared for TLM tests is shown in Fig. 9. The sample is at first mesa-etched usually to a

depth where there is a natural depletion or insulating layer. This is done in order to isolate

columns of the conductive epitaxial layer there by restricting current flow within the

column or the active device (in this case TLM) area. Metal contact pads, of finite width,

W, are then deposited on the mesa at a linearly increasing pad spacing, d, such that d1 <

d2 < d3. A constant current is passed between two adjacent pads through two probes; a

second set of probes are then used to measure the voltage drop enabling the total

resistance between the pads to be obtained. Separate current source and voltage senses

are preferred to cancel the resistance of the probes, which otherwise will compromise the

accuracy of the method especially when the measured total resistance is relatively low.

The process is repeated and the total resistance is plotted on a linear graph as a function

of pad spacing, d. An example is shown in Fig. 10.

21

Fig. 9 Schematic diagrams of metal contacts patterned with TLM structure on GaN

surface and measurement of I-V curves with four-point probes.

Fig. 10 Plot of total resistance as a function of TLM pad spacing.

22

The measured total resistance R(d), between two adjacent pads with spacing d, is

given by:

R(d) = 2Rc+Rsemi, (1)

where, Rc is the resistance due to the contact between metal and semiconductor, and Rsemi

is the resistance due to the semiconductor material.

Rsemi =Rsh*d/W, (2)

where Rsh is the sheet resistance of the semiconductor. Therefore,

R(d) = 2Rc + Rsh*d/W, (3)

Thus the contact resistance and the sheet resistance can be derived by plotting R(d)

vs d. The intercept of the R(d) vs d plot gives the contact resistance while the slope gives

the sheet resistance.

Rc = Intercept/2, (4)

Rsh = slope*W, (5)

The contact resistance depends on the size of the contact, therefore not a

normalized parameter to compare different samples. Therefore, specific contact

resistivity, ρc

ρc = Rc*Ac, (6)

where, Ac is the effective area of the contact, is chosen as the normalized parameter to

quantitatively evaluate and compare the quality of contacts. However, it is not the same

as the physical area of the contact metal, since the current does not flow uniformly in the

contact, as shown in Fig. 11. At the edge of the contact, the current flowing in (or out) is

significant. Moving away from that edge, the current drops off until, at the far edge, there

is no current[61]. This is known as “current crowding”. The transfer length, LT, is

23

introduced, which describes the average distance that an electron (or hole) travels in the

semiconductor beneath the contact before it flows up into or out of the contact. LT = 𝑅𝑐

𝑆𝑙𝑜𝑝𝑒,

is extracted from the standard plot shown in Fig. 10.

Therefore, the effective area of the contact can be treated as LT *W. Thus,

ρc = Rc* LT *W= 𝑅𝑐

2𝑊

𝑆𝑙𝑜𝑝𝑒, (7)

Therefore, both Rsh and ρc can be extracted from the linear plot of R(d) vs d.

Fig. 11 The current flow through the contacts and the semiconductor[61].

3.2 Fabrication of Contacts to GaN

The epitaxial layers of n-GaN and p-GaN were grown by Metal Organic Chemical

Vapor Deposition (MOCVD). The layer structure and doping profile is shown in Fig. 12.

For the n-GaN samples, an unintentionally doped c-plane GaN buffer layer (3.1 um) was

first grown on sapphire substrate, followed by 300 nm thick n-type GaN with a Si doping

concentration of 1×1017 cm-3. Then, the n-GaN structure was capped with 100 nm thick

heavily doped n-type GaN with a Si doping concentration of 1.7×1019 cm-3. For the p-

type contact studies,, an unintentionally doped c-plane GaN buffer layer (3.1 um) was

first grown on sapphire substrate, followed by 300 nm thick n-type GaN with a Si doping

24

concentration of 1×1017 cm-3 and 100 nm thick p-type GaN with a Mg doping

concentration of 3×1019 cm-3. The p-GaN structure was capped with 10 nm of p+GaN

with a Mg doping concentration of 1×1020 cm-3. The p-GaN samples were activated by

annealing at 700 °C in N2 ambient for 30 min.

Fig. 12 The layer structure and doping profile of the n-GaN and p-GaN samples used for

HT contact studies.

The determination of the specific contact resistivity (ρc) and sheet resistance (Rsh)

were carried out by the TLM measurement. TLM test structures were fabricated on both

n-GaN and p-GaN samples with layer structures as shown in Fig.12. Fig. 13 shows the

process flow of the contacts patterned with TLM structures. First, mesa regions for TLM

measurements were defined using inductively coupled plasma (ICP) etching to etch down

to the lightly doped n-type GaN. Next, TLM contact pads, with an area of 100×200 μm2,

were deposited by e-beam evaporation. A 1-min dip in HCl: DI water (1:3) mixture was

conducted to remove the native oxide layers prior to e-beam evaporation of the contact

pads. After deposition and lift-off of the metal contacts, we subjected the samples to a

rapid thermal annealing (RTA). The RTA conditions are different for each contact.

25

Further details about the RTA conditions have been discussed in the following two

chapters.

Fig. 13 Process flow of the contact patterned with TLM structures.

3.3 Contacts Characterization

The current-voltage (I-V) characteristics of the annealed contacts were measured

using a four-point probe station equipped with a Keithley 4200-SCS parameter analyzer.

The temperature dependence of ρc and Rsh were studied in the range from 25 °C to 400 °C,

varying the chuck temperature by using an Instec mK2000 temperature controller, as

shown in Fig. 14. To evaluate the stability of the contacts above 400 °C, a Minibrute

26

furnace, shown in Fig. 15, with N2 or air ambient was used for HT stress test. The contact

samples were measured before and after the HT stress to see if there is any degradation.

Fig. 14 Four-probe I-V characterization system with a thermal chuck which can heat the

sample up to 400 °C during the measurement in air.

27

Fig. 15 A Mini-brute furnace for HT stress with N2 and air flow.

To understand the mechanism that whether the ohmic contact is stable at high

temperature or it degrades due to the high temperature, high-resolution transmission

electron microscope (HR-TEM) was used to study the interface between metal and GaN.

The contacts samples were prepared by either focused ion beam (FIB) or wedge polishing

technique. Scanning transmission electron microscope (STEM) and in situ electron

energy loss spectroscopy (EELS) were employed to analyze the interfacial layer or phase

created by HT thermal stress. The electron microscopy provided strong explanation to

support the electrical characterization.

3.4 Summary

TLM structures were prepared to evaluate various contact metallurgy schemes.

Besides the electrical characterization done using TLM, the microscopic study of the

contacts was done with TEM, STEM and EELS. Electrical behavior along with the

28

microstructure images revealed all the information about the contacts and their

degradation mechanism. Chapters 4 and 5 describe the details of our study. Chapter 6

shows the application of the TLM results to derive the lifetime of the contacts.

29

Chapter 4

DEVELOPING OHMIC CONTACTS TO N-GAN FOR HT APPLICATIONS

4.1 Alloyed Ti/Al/Ni/Au Contacts

The Ti/Al/Ni/Au multilayer (of thicknesses - 20nm, 120nm, 30nm, and 50nm,

respectively) metal contacts to n-GaN show excellent ohmic behavior after RTA at

800 °C in N2 for 30s, as was discussed in Chapter 2. We first measured the TLM

structures at room temperature in air. The I-V curves, shown in Fig. 16 (a), are quite

linear, resulting in the plot of measured resistance versus TLM contact spacing, as shown

in Fig. 16 (b). A sheet resistance, Rsh, of 288 Ω/□ and the specific contact resistivity, ρc,

of 6.8×10-6 Ω cm2 was derived from the TLM data calculated using the method

introduced in Chapter 3.

Fig. 16 Electrical characteristics of the alloyed Ti/Al/Ni/Au contacts to n-GaN measured

at room temperature in air, (a) I-V curves measured between adjacent TLM pads with

spacing increasing from 3um to 30um, and (b) plot of measured resistance versus TLM

contact spacing.

30

To understand the HT behavior, the sample was heated up to 400 °C in air and the

TLM measurements were taken at 300 °C and 400 °C, respectively. The temperature

dependent I-V characteristics of the alloyed Ti/Al/Ni/Au contacts, measured between two

adjacent pads with a spacing of 10um, is shown in Fig. 17. As the measurement

temperature increased, the I-V curves remained linear, but the current reduced slightly,

indicating an increase of the measured resistance. Fig. 18 shows the temperature

dependent contact resistivity and sheet resistance of the alloyed Ti/Al/Ni/Au to n-GaN

contacts. ρc, ~6×10-6 Ω cm2, is indeed independent of the measurement temperature,

while Rsh increases significantly as the measurement temperature increases, from 288 Ω/□

at room temperature to 370 Ω/□ at 400 °C. The increase of n-GaN sheet resistance is

most likely due to the decrease of electron mobility at high temperature. Therefore,

although the contacts resistivity of the alloyed Ti/Al/Ni/Au to n-GaN contacts did not

change with the ambient temperature, the n-GaN layers became more resistive as the

temperature increased.

31

Fig. 17 Temperature dependent I-V characteristics of the alloyed Ti/Al/Ni/Au contacts,

measured between two adjacent pads with gap spacing 10um.

Fig. 18 Contact resistivity and sheet resistance of the alloyed Ti/Al/Ni/Au to n-GaN

contacts as a function of measurement temperature.

-0.4 -0.2 0.0 0.2 0.4-0.03

-0.02

-0.01

0.00

0.01

0.02

0.03

Curr

ent (A

)

Voltage (V)

RT

300oC

400oC

0 100 200 300 4001E-6

1E-5

1E-4

Contact resistivity

Sheet resistance

Temperature (C)

c (cm

2)

RT

280

320

360

Rsh (

)

32

To further investigate the thermal stability of the alloyed Ti/Al/Ni/Au to n-GaN

contacts at 400 °C in air, the samples were held at 400 °C for 3h, during which the TLM

structures were measured multiple times. Fig. 19 shows the time evolution of contact

resistivity and sheet resistance as the samples were held at 400 °C for 3 h. Both ρc and Rsh

were very consistent at 400 °C and no degradation could be observed during the 3h high

temperature testing. When the sample was cooled down to room temperature, both ρc and

Rsh returned to their initial values measured prior to the high temperature testing. Based

on the above analysis, the sheet resistance of n-GaN is a function of measurement

temperature, but it is independent of the time subjected at those high temperatures. The

change in sheet resistance was reversible when the sample was cooled down to room

temperature. The contact resistivity is independent of the measurement temperature and it

was very steady at 400 °C with ambient air.

Fig. 19 Contact resistivity and sheet resistance of the Ti/Al/Ni/Au to n-GaN contacts as a

function of time under 400 °C.

0 60 120 1801E-6

1E-5

1E-4

c (cm

2)

Contact resistivity

Sheet resistance

Time (min)

280

300

320

340

360

380

400

Rsh (

)

RTRT 400C

33

To test the stability of the alloyed Ti/Al/Ni/Au to n-GaN contacts at 450 °C in air,

the sample was subjected to a Minibrute furnace and baked at 450 °C for 8 hours. Fig. 20

(a) and (b) shows the comparison of contacts surface before and after the 8-hour bake at

450 °C in air, respectively. No visible difference was observed. Also, there was

negligible change in contact resistivity and sheet resistance that were extracted from the

plots of measured resistance versus TLM contact pads spacing, pre-bake (c) and post-

bake (d).

34

Fig. 20 Microscopy images of the alloyed Ti/Al/Ni/Au contacts surface, pre (a) and post

(b) the 450 °C bake, respectively. Plots of measured resistance versus TLM contact pads

spacing, pre (c) and post (d) the 450 °C bake, respectively.

To find out the temperature limit of the contact in air ambient, a new sample was

prepared. The sample with the alloyed contacts after post-deposition RTA offered a ρc of

5.9×10-6 Ω cm2 with a Rsh of 441 Ω/□. Then the sample was sequentially subjected to

400 °C, 500 °C, 550 °C, and 600 °C in the Minibrute furnace with air ambient for 4 hours

at each temperature. The sample was taken out and cooled down to RT for the TLM

measurement after the thermal stress at each temperature. The change of ρc and Rsh are

shown in Fig. 21. The sheet resistance was very consistent after each thermal stress,

35

showing values around 445 Ω/□, indicating thermal stress up to 600 °C did not cause

significant change in the conduction layer of n-GaN. The contact resistivity increased

after thermal stress at 500 °C for 4 hours, but it then decreased after thermal stress at

550 °C for 4 hours. Finally, ρc showed a tendency to stabilized around 1×10-6 Ω cm2 after

thermal stress at 600 °C for 4 hours.

Fig. 21 The change of contact resistivity and sheet resistance of the alloyed Ti/Al/Ni/Au

contacts after thermal stress at 400 °C, 500 °C, 550 °C, and 600 °C for 4 hours each with

ambient air.

Similar study was done on the Ti/Al/Mo/Au contacts, which uses a layer of Mo

instead of Ni for the diffusion barrier. The Ti/Al/Mo/Au contacts, with thicknesses of

20nm, 120nm, 30nm, and 50nm, respectively, were also subject to RTA at 800 °C in N2

for 30s. The TLM test structures showed good ohmic behavior with ρc of 3.7×10-5 Ω cm2

and Rsh of 314 Ω/□. Fig. 22 shows the change of ρc and Rsh of the alloyed Ti/Al/Mo/Au


Initial 350 400 450 500 550 6001E-7

1E-6

1E-5

1E-4

c(c

m2)

Thermal Stress Temperature (C)

300

350

400

450

500

Rsh (

)

36

ambient air. Similar to those with the alloyed Ti/Al/Ni/Au contacts, sheet resistance was

not affected by the thermal stress at temperatures up to 600 °C while contact resistivity

increased after the 500 °C thermal stress and decreased after the 550 °C thermal stress.

Fig. 22 The change of contact resistivity and sheet resistance of the alloyed Ti/Al/Mo/Au


ambient air.

In summary, the Ti/Al/Ni/Au contacts as well as the Ti/Al/Mo/Au contacts

showed very low contact resistivity and very good stability after the post-deposition RTA

at 800 °C. High temperature stress, up to 600 °C, did not affect the sheet resistance of n-

GaN. 500 °C thermal stress would increase the contact resistivity, but higher temperature

(> 550 °C) stress tends to improve the contact resistivity.

To understand the underlying reason for the high temperature stability of the

alloyed contacts, transmission electron microscope (TEM) was used to study the

microstructure of the contacts interface. Fig. 23 shows the TEM images of the alloyed

Initial 350 400 450 500 550 6001E-6

1E-5

1E-4

1E-3

Rsh (

)

c(c

m2)


100

200

300

400

37

Ti/Al/Ni/Au contacts to n-GaN. Clearly there is a thin layer between the GaN and the

metal alloy, shown in Fig. 23 (a). This layer is identified as TiN because it has a cubic

crystal structure and shows a lattice constant of 4.23 Å, shown in Fig. 23 (b), matching

that of TiN. The layer of TiN was formed by the reaction between Ti and the N atoms

from GaN lattice during the RTA at 800 °C. As mentioned in Chapter 2, the formation of

TiN, creating a lot of N vacancies, is responsible for the low contact resistivity of the

alloyed contacts. On the other hand, TiN is a hard, dense, refractory material with high

electrical conductivity[62], [63]. It has a melting point of 2,930 °C and it won’t oxidize

until 800 °C[63]. The good corrosion and erosion resistance of TiN, and its relative

inertness make it widely used as diffusion barriers in microelectronic devices. Therefore,

the layer of TiN is the key for the stability of the alloyed contacts at high temperature,

very stable itself and preventing further interfusion and reaction between metal and GaN.

Fig. 23 Cross-section TEM images of the alloyed Ti/Al/Ni/Au contacts to n-GaN. (a)

Low magnification image of the interface and (b) high resolution image of the interfacial

TiN layer.

38

4.2 Non-Alloyed Al/Au Contacts

In the previous discussion, the alloyed Ti/Al/Ni/Au contacts showed very low

contact resistivity, ρc ~6×10-6 Ω cm2, and very good stability up to 600 °C in air.

However, the contact surface is very rough, with root-mean-square (rms) roughness of 70

nm, probably due to the low melting point of Al, causing the lateral diffusion of Al and

ball up. Also the alloyed contacts have several complex phases, e.g. Al-Au, Al-Ni, Ti-Al

and TiN, which could also increase the surface roughness. Besides, the interface of metal

and GaN is relatively rough, shown in Fig. 23(a), due to the reaction between metal and

GaN during the 800 °C annealing. In addition, Fig. 23 (b) shows that there is a thin layer

with thickness of 5nm-10nm “damaged” by the alloying, giving the possibility of metal

diffusion or spiking into GaN surface. The diffusion of metal into GaN could kill some

GaN devices in which the surface layer is very thin, only a few nanometers. Therefore,

the non-alloy ohmic contacts is also necessary for GaN high temperature applications.

As discussed in Chapter 2, Al forms ohmic contact to n-GaN without annealing.

Adding a cap layer of Au could help to prevent the oxidation of Al and increase the metal

conductivity as well. The Al/Au bilayer contacts with thickness of 30nm and 300nm,

respectively, were tested at room temperature. Fig. 24(a) shows the I-V curves measured

on the TLM structures and Fig. 24(b) shows the plot of measured resistance versus TLM

contact pad spacing. The extracted Rsh and ρc were 548 Ω/□ and 3×10-4 Ω cm2,

respectively. Although the contact resistivity is much higher than that of the alloyed

contacts, it is low enough for the purpose of photovoltaic applications.

39

Fig. 24 Electrical characteristics of the as deposited Al/Au contacts to n-GaN measured at

room temperature in air, (a) I-V curves measured between adjacent TLM pads with

spacing increasing from 3um to 30um, and (b) plot of measured resistance versus TLM

contact spacing.

To study the high temperature stability and limit of the Al/Au ohmic contacts, the

contact sample was subjected to 400 °C, 500 °C, 550 °C, and 600 °C, sequentially, in the

Minibrute furnace with air ambient for 4 hours at each temperature. The change of ρc and

Rsh is shown in Fig. 25. Rsh increased from 548 Ω/□ to 600 Ω/□ after thermal stress at

400 °C for 4 hours, indicating possible reaction between the metal and n-GaN surface

layer, but it did not increase further after higher temperature stress. ρc was reduced

significantly by the 400 °C thermal stress, from 3×10-4 Ω cm2 to 1.6×10-5 Ω cm2. Higher

temperature stress caused slight reduction in ρc. The Rsh and ρc measured after thermal

stress at 600 °C for 4 hours are 601 Ω/□ and 5×10-6 Ω cm2, respectively. Therefore, the

non-alloyed Al/Au contacts also show very low contact resistivity and very good stability

for 600 °C operation.

40

Fig. 25 The change of contact resistivity and sheet resistance of the Al/Au contacts after

thermal stress at 400 °C, 500 °C, 550 °C, and 600 °C for 4 hours each with ambient air.

To understand the underlying mechanism of forming such contacts with low

resistivity and high temperature stability, TEM study was conducted on the Al/Au sample

after the 4 hour thermal stress at 600 °C in air. The cross-section view of metal-GaN

interface is shown in Fig. 26. The TEM image indicates that there is a metal layer with

thickness around 66nm on the interface and another layer with thickness around 237nm

on top of that, while the as deposited thickness of Al and Au are 30nm and 300nm,

respectively. Scanning transmission electron microscope (STEM) and in situ electron

energy loss spectroscopy (EELS) were used to identify these two layers. Fig. 27 shows a

high resolution STEM image of the metal-GaN interfacial area. A flat and sharp interface

is observed, indicating that the reaction between metal and GaN and the diffusion of

Initial 350 400 450 500 550 6001E-6

1E-5

1E-4

1E-3

c(c

m2)


500

550

600R

sh (

)

41

metal into GaN are negligible. EELS scan was conducted across the 66nm interfacial

layer, shown in Fig. 28(a). The EELS spectrum of N, O, Ga, Al, and Au are shown in Fig.

28(b), clearly indicating that the 66 nm interfacial metal layer consists of Al, O , and Au,

while the top layer consists of only Au. This suggests that there was gold diffusion into

the interface and oxygen reaction with aluminum during the thermal stress in air. It is

interesting that inspite of the presence of oxygen with Au, which has high work function,

the the interfacial layer complex allows the formation of low resistivity contacts to n-

GaN. Further study of the Al-O-Au complex in forming ohmic contact is needed.

However, it seems that this Al-O-Au complex reaches a saturation state, which is very

stable at high temperature.

Fig. 26 Cross-section TEM image of the Al/Au contacts to n-GaN after thermal stress at

600 °C in air for 4 hours.

42

Fig. 27 STEM image of the interfacial area on the Al/Au to n-GaN contacts after thermal

stress at 600 °C in air for 4 hours.

Fig. 28 (a) EELS scan trace, across the 66 nm interfacial layer and (b) EELS spectrum of

N, O, Ga, Al, and Au on the Al/Au to n-GaN sample after thermal stress at 600 °C in air

for 4 hours.

5 nm5 nm

43

4.3 Summary

In this chapter, both alloyed and non-alloyed contacts to n-GaN were developed

for high temperature applications. Both of them show low resistivity, ~10-6 Ω cm2, and

good stability up to 600 °C in air.

The alloyed Ti/Al/Ni/Au contact can be easily achieved, showing ρc ~ 6.8×10-6 Ω

cm2 at room temperature. Temperature dependent measurements, up to 400 °C, show that

the contact resistivity is independent of the measurement temperature while the sheet

resistance of n-GaN increased from 288 Ω/□ at room temperature to 370 Ω/□ at 400 °C,

due to the decrease of electron mobility at high temperature. It must be mentioned that

the change in the sheet resistance of n-GaN was reversible. After the contacts were

cooled down to room temperature the contact resistivity restored its initial value. Stability

of the contact was first tested by subjecting the sample to 450 °C for 8 hours. No

degradation in contact resistivity and sheet resistance was observed after the thermal

stress. The contact was tested further at higher temperature up to 600 °C. No change in

sheet resistance of n-GaN was observed, and in fact the contact resistivity improved

slightly after the contacts were subjected to 600 °C for 4 hours in air. The alloyed

Ti/Al/Mo/Au contact showed similar properties to the alloyed Ti/Al/Ni/Au contact, but

offered higher contact resistivity. TEM study of the contact microstructure indicates that

the formation of TiN on the metal-GaN interface is responsible for the low resistivity and

high stability contact.

The non-alloyed Al/Au contact offered a ρc ~ 3×10-4 Ω cm2 before any high

temperature testing was conducted. The contact resistivity decreased after thermal stress

at 400 °C, 500 °C, 550 °C, and finally stabilized to 5×10-6 Ω cm2, same as that of the

44

alloyed contact, after thermal stress at 600 °C for 4 hours in air. Instead of showing

degradation, ρc was reduced by two orders of magnitude. TEM, STEM, and EELS studies

showed the formation of Al-O-Au complex at the interface, caused by the thermal stress

in air. The interfacial Al-O-Au played an important role in reducing the contact resistivity

and maintaining good stability at temperatures up to 600 °C in air.

45

Chapter 5

DEVELOPING OHMIC CONTACTS TO P-GAN FOR HT APPLICATIONS

5.1 Annealed Ni/Au Contacts to p-GaN

5.1.1 Electrical Characterization

The most commonly used metallization structure for p-GaN ohmic contacts has

been the Ni/Au stack. In fact, such metallization scheme was used even in the early

generations of GaN blue LEDs as early as 1993[64].

After the deposition and lift-off of the Ni/Au layers of thickness 20 nm/200 nm,

the samples were subjected to a rapid thermal annealing (RTA) process at 500 °C, 550 °C,

and 600 °C in a N2 ambient for various time intervals. The current-voltage (I-V)

characteristics of the annealed contacts were measured using a four point probe station

equipped with a Keithley 4200-SCS parameter analyzer. Fig. 29 shows the extracted

contact resistivity as a function of RTA time at 500 °C, 550 °C, and 600 °C, respectively.

The contacts formed by RTA at 600 °C for 3 minutes showed the lowest specific contact

resistivity ρc, 1.6×10-3 Ω•cm2, with a sheet resistance Rsh of 1.3×105 Ω/□. ρc and Rsh were

extracted from the linear plot of measured resistance (Rmeasured) versus contact pad

spacing, shown in Fig. 30, according to the TLM model. The linear I-V curves of this

optimized contacts are shown in the graph inset of Fig. 30, depicting very good ohmic

behavior at room temperature (25 °C). This optimized Ni/Au contact were then subjected

to the study for high temperature applications.

46

Fig. 29 Extracted contact resistivity as a function of rapid thermal annealing time at 500

°C, 550 °C, and 600 °C, respectively.

Fig. 30 Measured resistance versus TLM pad spacing after the Ni/Au to p-GaN contacts

were annealed at 600 °C in N2 for 3 min. The inset picture shows the I-V curves

measured between adjacent contact pads with various spacing.

0 5 10 15 201E-3

0.01

0.1

c(c

m2)

RTA time (min)

500 °C

550 °C

600 °C

47

A. Dependence of ρc and Rsh on temperatures up to 390 °C

The temperature dependence of the specific contact resistivity and sheet resistance

were studied in the range of 25-390 °C, varying the temperature of the probe station

chuck by using an Instec mK2000 temperature controller. Fig. 31 shows the I-V

characteristics of the contacts, measured at various temperatures between 25 °C and 390

°C, taken between two adjacent TLM pads with gap spacing 5 um. The contacts remain

ohmic at all temperatures from 25 °C to 390 °C with the current increasing with the

increasing temperature. The current at 390 °C was 10 times higher than that at 25 °C. The

ρc and Rsh values of the contacts, extracted from the TLM measurements, as a function of

the measurement temperature are shown in Fig. 32 (a) and (b), respectively. Both specific

contact resistivity and sheet resistance decreased with increasing measurement

temperature. ρc decreased by a factor of 10, i.e., from 1.6×10-3 Ω•cm2 at 25 °C to 1.6×10-4

Ω•cm2 at 390 °C. Rsh decreased by a factor of 11, i.e., from 1.3×105 Ω/□ at 25 °C to

1.2×104 Ω/□ at 390 °C. The reduction in ρc and Rsh was due to the increased hole

concentration at high temperature, as expected [65], leading to 10 times increase in the

current. Therefore, although ohmic contacts to p-GaN is difficult to achieve at RT, the

Ni/Au to p-GaN contacts shows superior ohmic behavior at elevated temperatures.

48

Fig. 31 Temperature dependent I-V characteristics of the annealed Ni/Au to p-GaN

contacts, taken between two adjacent pads with gap spacing 5um.

49

Fig. 32 Temperature dependent specific contact resistivity ρc, (a), and sheet resistance

Rsh, (b) of the annealed Ni/Au to p-GaN contacts.

50

B. Stability of the contacts at 450 °C in air

The thermal stability of the annealed Ni/Au to p-GaN contacts was tested on several

samples with the optimized annealing condition, showing an average initial ρc of 2.2×10-3

Ω•cm2. The samples were placed in a Minibrute furnace at 450 °C with ambient air for a

maximum time of 48 hours. During this 48 hours of test the samples were taken out of the

furnace and the I-V characteristics were measured at room temperature at regular

intervals. Fig. 33 shows the plot of measured resistance versus contact pad spacing after

the sample was subjected to 450 °C in the ambient of air for 48 hours. The I-V curves

obtained after 48 hours of high temperature thermal stress, as shown in the inset of Fig.

33, validate that the contacts maintained very good ohmic behavior. The ρc and Rsh

extracted from Fig. 33 are 1.4×10-3 Ω·cm2 and 1.32×105 Ω/□, respectively. The reduction

in ρc and Rsh, compared with those before the high temperature stress, are approximately -

36% and 6%. While the change of Rsh was negligible while the improvement in the ρc was

possibly a result of further annealing of the contacts, leaving some room for typical errors

incurred in the TLM method [66].

51

Fig. 33 Measured resistance versus TLM pad spacing after the annealed Ni/Au contacts

samples were subjected to 450 °C in air for 48 h. The inset picture shows the I-V curves

measured between contact pads with various spacing from 3um to 30 um.

Fig. 34 (a) shows the time evolution of the specific contact resistivity measured and

averaged over at least four samples after thermal stress at 450 °C in air for up to 48 hours.

A significant increase and then decrease in ρc was observed during the initial 10-hour

“burn-in” period[41]. After the first 10 hours, the ρc decreased slowly and stabilized at

~2×10-3 Ω·cm2. No degradation was observed thereby demonstrating excellent thermal

stability of these contacts at 450 °C in air. Fig. 34 (b) shows the time evolution of the

sheet resistance during the 48-hour thermal stress at 450 °C in air, obtained from the

same TLM readings. The change in Rsh is negligible, as expected, because the both

52

metals, Ni and Au, don’t react with GaN and 450 °C is far below the dissociation

temperature limit of GaN[2], [67]. The measurements were repeated for multiple TLM

structures across the samples and the results were consistent, showing excellent stability

of our contact process.

53

Fig. 34 Time evolution of the specific contact resistivity ρc, (a), and sheet resistance Rsh,

(b), of the annealed Ni/Au contacts over the 48-hour thermal stability test at 450 °C in

air.

54

C. Contacts degradation above 450 °C

To identify the high temperature limit of the contacts, the test sample was

sequentially subjected to thermal stress at 500, 550, 600, 650, and 700 °C for 4 hours

each in the Minibrute furnace with ambient air. I-V characteristics were measured at room

temperature after each thermal stress and the specific contact resistivity was extracted.

Fig. 35 (a) shows the I-V characteristics of the contacts between two adjacent TLM pads

with a spacing of 5 μm, at room temperature and then before and after each thermal stress

cycle at temperatures up to 650 °C. The I-V curves remain linear even after thermal stress

at 650 °C for 4 hours, but the current is significantly reduced as the temperature increases

to 650 °C. Fig. 35 (b) shows the evolution of the specific contact resistivity caused by

the thermal stress at temperatures above 450 °C. ρc increases by 69% after 4 hours of

thermal stress at 500 °C, further increasing by 400% after 4 hours of thermal stress at 550

°C, by 180% after 4 hours of thermal stress at 600 °C, and by 152% after 4 hours of

thermal stress at 650 °C. The sheet resistance of p-GaN also increased significantly

especially after thermal stress at 600 °C and 650 °C, shown in Fig. 36, indicating that the

hole concentration at the surface was reduced probably caused by high temperature

thermal stress. The reduction in the surface hole concentration is responsible for the

degradation in contact resistivity. The contacts showed rectifying I-V curves after being

subjected to 700 °C in air for 4 hours and the current was reduced by three orders of

magnitude compared to the initial value, shown in Fig. 37.

55

Fig. 35 (a) I-V characteristics measured at room temperature and taken between two

adjacent TLM pads with gap spacing 5 um and (b) the extracted ρc of the Ni/Au contacts

after being subjected to thermal stress at 500 °C, 550 °C, 600 °C, and 650 °C in series in

air.

56

Fig. 36 The extracted Rsh of the Ni/Au contacts after being subjected to thermal stress at

500 °C, 550 °C, 600 °C, and 650 °C in series in air.

Fig. 37 I-V characteristics of the Ni/Au contacts measured at room temperature after

being subjected to 700 °C for 4 hours in air.

Initial 500 550 600 6501.2x10

5

1.4x105

1.6x105

1.8x105

2.0x105

Rsh(

/)


After 4hAfter 4h

After 4h

After 4h

-0.4 -0.2 0.0 0.2 0.4-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

3m

5m

10m

15m

20m

25m

30m

I (

A)

Voltage (V)

57

5.1.2 Microstructure Characterization

In order to understand the root cause for the good stability of the annealed Ni/Au

contacts at high temperatures up to 450 °C in air, we performed transmission electron

microscopy (TEM) study on the contact interface. The first set of TEM was conducted

after the sample was annealed, but before any further thermal stress.

The cross-section image of the annealed contacts is shown in Fig. 38. Clearly it

has three layers and two interfaces, marked as interface 1 and interface 2. The high-

resolution TEM images of interface 1 and interface 2 are shown in Fig. 39 and Fig. 40,

respectively. The three layers are identified as NiO, Au, and GaN by the lattice

parameters measured from the high-resolution TEM images, shown in Fig. 39 and Fig. 40.

The TEM study provides direct evidence of the layer reversal between the as deposited

Ni and Au. Au diffuses into the metal-GaN interface while Ni diffuses out to the surface

of Au and reacts with air to form NiO during the post-deposition annealing. In addition,

Fig. 40 shows very flat and sharp interface between Au layer and GaN layer, with Au

lattice grown epitaxially on GaN lattice. The epitaxial structure of Au-GaN interface is

believed to play an important role in forming good ohmic contacts to p-GaN. Early

reports show that direct deposition of Au on p-GaN surface does not result in low-

resistance ohmic contacts [68], [25]. It must be mentioned here that the deposition of Au,

as a noble metal, will start with nucleation of gold particles, followed by lateral cluster

growth, then coarsening, and finally a continuous layer [69]. Therefore, the Au layer

would be polycrystalline or non-crystalline, and it won’t form epitaxial structure with

GaN. However, in the Ni/Au contacts system, a wetting layer of Ni is used to remove the

surface oxide by reacting with O and diffusing out to the surface of Au while Au diffuses

58

into the metal-GaN interface and form intimate epitaxial structure with GaN lattice. The

epitaxial structure of Au on p-GaN is also responsible for the high temperature stability

of the contacts because Au doesn’t react with GaN and Au doesn’t diffuse into GaN

unless the temperature is very high.

Fig. 38 Cross-section TEM image of the Ni/Au contacts after post-deposition annealing,

showing three layers with two interfaces.

59

Fig. 39 High-resolution TEM image of the top interface, marked as interface 1, of the

annealed Ni/Au contacts.

Fig. 40 High-resolution TEM image of the metal-GaN interface, marked as interface 2, of

the annealed Ni/Au contacts.

60

The samples with the TLM test structures were subjected to thermal stress at

700 °C for 4 hours in air, showing rectifying I-V characteristics, was also prepared for

TEM analysis to investigate the degradation mechanism.

The cross-section TEM image of the degraded contacts is shown in Fig. 41. Both

the metal-GaN interface and GaN surface are very rough with a new layer formed

between GaN and metal, indicating strong reaction on the interface during the thermal

stress at 700 °C in air. The high resolution TEM image for the interfacial area is shown in

Fig. 42. Au and GaN are identified by the lattice parameters and the interfacial layer

shows similar lattice parameter to GaN, implying the formation of Gallium oxide. The

insulating gallium oxide layer can be produced by thermal oxidation of Gallium-polar

GaN at high temperature [70]–[73]. Therefore, the formation of Gallium oxide on the

contacts interface is believed to be the root cause for the degradation in contact resistivity.

The study of the degradation mechanism provided us with evidences of oxidation.

Extension of the lifetime or increase of the operating temperature limit should be possible

by encapsulation of the GaN surface with SiN with appropriate wire bonding techniques,

so that there is no path for the oxygen to react with GaN.

61

Fig. 41 Cross-section TEM image of the Ni/Au contacts after thermal stress at 700 °C for

4 hours in air.

Fig. 42 High-resolution TEM image of the interfacial area on the Ni/Au contacts after

thermal stress at 700 °C for 4 hours in air.

62

5.2 Indium Tin Oxide Contacts to p-GaN

Previous section shows that the annealed Ni/Au to p-GaN contact is very stable at

temperatures up to 450 °C in air and the contact resistivity increases only slightly after 4

hours at 500 °C. However, the low transmittance associated with this metal contact, with

total thickness of 220 nm, acts as a bottleneck when applied to optoelectronics, e.g. LED

and solar cell. Although the transmittance can be increased by using very thin metal layer,

it leads to large series resistance effects that in turn cause lower efficiency, increased

device temperature, non-uniform light emission, and so on. Transparent and conducting

contacts can be used to combat this problem. Indium tin oxide (ITO) has been widely

used as the transparent conductive electrodes in organic and inorganic optoelectronic

devices because of its high optical transmittance in the visible wavelength and good

electrical conductivity [74]–[76]. In commercial GaN-base LEDs and solar cells, ITO is

deposited on p-GaN is used as the transparent conductive electrodes to spread the current

more uniformly, and many studies on the ITO ohmic contacts to p-GaN have been

reported [77]–[80].

In addition to the contact resistivity between ITO and p-GaN, the film resistivity

of ITO is another factor affecting the series resistance of InGaN solar cell, because ITO is

a semiconductor with typical resistivity two orders of magnitude higher than that of metal,

so the film resistivity is not negligible. The conductivity of ITO is attributed to oxygen

vacancies in the film, substantial tin donor dopants, and film crystallinity [81]–[83].

Therefore, the resistivity is highly dependent to the growth condition, annealing

temperature, and oxygen pressure. In the high temperature solar cell, the resistivity of

ITO film could be affected by the operating temperature and ambient.

63

The ITO contact used for high temperature InGaN solar cell was prepared by the

company of Photonitride Devices Inc. The ITO film with thickness of 150 nm was

deposited in vacuum by e-beam evaporation using a source composed of 90 wt% In2O3

and 10 wt% SnO2. Since the as-deposited ITO film is not transparent due to oxygen

deficiency, annealing at 500 °C in air for 1 hour is needed to turn the opaque film into a

transparent film. Two TLM structures, A and B shown in Fig. 43 (a) and (b), respectively,

were made to test the contact resistivity and film resistivity. Measurements on TLM A

gave the contact resistivity of ITO to p-GaN contact while measurements on TLM B gave

the sheet resistance of ITO layer.

Fig. 43 Two TLM structures to test the contact resistivity of ITO to p-GaN, (a), and the

sheet resistance of ITO film, (b), respectively.

64

Fig. 44 (a) shows the I-V characteristics of the ITO to p-GaN contact after post-

deposition annealing at 500 °C in air for 1h, measured on TLM A, depicting good ohmic

behavior of the contact. The plot of measured resistances versus TLM contact spacing,

shown in Fig. 44 (b), gives ρc of 9.4×10-3 Ω·cm2. The sheet resistance of ITO film,

extracted from measurements on TLM B, is 27.8 Ω/. The converted resistivity of ITO

is 4.2×10-4 Ω·cm, showing very good conductivity.

Fig. 44 (a) I-V characteristics of the ITO to p-GaN contact after post-deposition

annealing at 500 °C in air for 1h, measured on TLM A, and (b) plot of measured

resistance versus TLM contact spacing.

65

The high temperature stress tests were performed in a N2 ambient first, because

the high temperature solar cell was preferred to operating in a non-oxidizing ambient.

The annealed samples were subjected to the Mini-brute furnace at temperatures up to

450 °C with N2 flow, which was not a pure N2 ambient.

Fig. 45 (a) shows the I-V characteristics measured on TLM A of the ITO to p-

GaN contact after thermal stress at 450 °C in N2 for 8h,. The currents were significantly

reduced compared with those before the thermal stress, indicating the degradation of the

contacts. The plot of measured resistance versus TLM contact spacing, shown in Fig. 45

(b), gives ρc of 2.7×10-2 Ω·cm2. The sheet resistance of ITO film, extracted from

measurements on TLM B, is 11.5 Ω/, much lower than that before the thermal stress at

450 °C in N2. The reduction in the sheet resistance is possibly due to the out-diffusion of

oxygen in the ITO film during the thermal stress in N2 ambient, which would increase the

carrier concentration by introducing more oxygen vacancies, and thus the ITO film

becomes more conductive. The out-diffusion of oxygen is most likely the cause for the

shift of ITO work function. ITO was reported to have work functions between 4.8 and 5

eV, which can be controlled by altering the chemical composition[84]. The work function

of ITO decreases as the carrier concentration increases[85], [86]. Out-diffusion of oxygen

will reduce the work function of ITO, while absorption of oxygen will increase the work

function. The decrease in the work function of ITO, caused by the thermal stress in N2

ambient, is believed to be responsible for the contact degradation. To take a close look at

this phenomenon, thermal stress test at lower temperatures in N2 were performed. No

degradation was observed after a 8h thermal stress at 300 °C in N2, however, ρc increased

66

by 25% when the thermal stress was conducted at 350 °C in N2 for 8h. Therefore, the out-

diffusion of oxygen happens at temperature around 350 °C in non-oxygen ambient.

Fig. 45 (a) I-V characteristics of the ITO to p-GaN contact after thermal stress at 450 °C

in N2 for 8h, measured on TLM A, and (b) plot of measured resistance versus TLM

contact spacing.

67

The above discussion indicates that the ITO to p-GaN contact is stable up to

300 °C in N2 ambient. Higher temperature stress in non-oxygen ambient will cause out-

diffusion of oxygen, reducing the work function of ITO which in turn increases the

contact barrier. To solve this problem, thermal stress tests were conducted in air as well

carrying plentiful oxygen supply. Fig. 46 (a) shows the I-V characteristics of the ITO to

p-GaN contact after thermal stress at 450 °C in air for 8h, measured on TLM A are very

similar to those before the thermal stress test. The ρc extracted from the plot of measured

resistance versus TLM contact spacing, shown in Fig. 46 (b), is 9.2×10-3 Ω·cm2. The

sheet resistance of ITO film, extracted from measurements on TLM B, is 28.9 Ω/. The

difference in those values, before and after thermal stress, is negligible, indicating good

stability of the ITO contact at temperatures up to 450 °C in air.

The sample was then thermally stressed at 500 °C in air for 8 hours. The currents

measured on the TLM A were significantly reduced, shown in Fig. 47 (a). The extracted

ρc was 3.6×10-2 Ω·cm2, shown in Fig. 47 (b), three times higher than that after the 450 °C

thermal stress in air. The sheet resistance of ITO film, extracted from measurements on

TLM B, also increased significantly from 28.9 Ω/ to 48 Ω/. The increase in sheet

resistance is probably due to the reduction of oxygen vacancy. Thermal stress at high

temperature in air will introduce more oxygen into ITO, so the oxygen deficiency can be

recovered. In addition, the donor dopant Sn in ITO may also form Sn-O bonds, reducing

the concentration of Sn4+. Lower concentration of Sn4+ results in higher ITO resistivity

[87]. Therefore, in contrast to the degradation in N2 ambient caused by the out-diffusion

of oxygen, the degradation in air at high temperature is attributed to the absorption of

excessive oxygen.

68


in air for 8h, measured on TLM A, and (b) plot of measured resistance versus TLM

contact spacing.

69


in air for 8h, measured on TLM A, and (b) plot of measured resistance versus TLM

contact spacing.

70

5.3 Summary

In this chapter, two types of contacts for p-GaN have been extensively studied,

viz. annealed Ni/Au contact and annealed ITO contact.

The annealed Ni/Au contacts, measuring an initial ρc ~2×10-3 Ω·cm2 at room

temperature retained its ohmic behavior at all temperatures from 25 °C to 390 °C. It

showed superior ohmic behavior at elevated temperatures. Both contact resistivity and the

sheet resistance of p-GaN reduced by one order of magnitude at 390 °C, and expected to

reduce further at 450 °C, due to the increased hole concentration at high temperature. The

contact showed very good stability at 450 °C in air with no degradation observed after 48

hours. TEM study showed that the formation of Au/p-GaN epitaxial structure after

annealing allowed for the low contact resistivity and good stability at high temperature as

well. Our studies suggested that the observed degradation at higher temperature is

possibly caused by the oxidation of GaN surface. The TEM study confirmed that there

was a layer of Ga2O3 formed on the interface between metal and GaN after the contact

being subjected to 700 °C in air for 4 hours. Therefore, the annealed Ni/Au contacts is

good for stable operation at temperatures up to 450 °C in air.

The annealed ITO contact with ρc ~1×10-2 Ω·cm2, is stable up to 300 °C in N2.

Degradation in ITO contacts at higher temperature is probably caused by the out-

diffusion of oxygen from ITO, resulting in the decrease of ITO work function. For up to

450 °C the contact was observed to be stable in air. The degradation in air at higher

temperature in air is likely due to the absorption of excessive oxygen in the ITO film,

reducing the conductivity of the contact.

71

Chapter 6

LONG TERM RELIABILITY AT HIGH TEMPERATURE

6.1 Arrhenius Model for Accelerated Life Testing

In order to produce and continually improve the reliability of electronic devices, it

is necessary to find a way of predicting the lifetime of a device. The estimation of devices

lifetime at a designated temperature is important to both manufacturers and users.

Lifetime testing is a useful method for achieving three goals, namely, selecting durable

devices, estimating the median lifetime of the device, and finding the optimal condition

for operation. For example, we would like to know how long would the contacts last

before significant degradation happens at 450°C operation. The answer to this question

can be answered through statistical analysis of actual lifetime data. However, the

lifetimes could be thousands of hours or even longer, so it is generally not practical to test

the devices at normal operation conditions. Some methods must be utilized in order to

obtain the failure characteristics in a relatively reasonable amount of time. The most

prevalent lifetime prediction method used extensively in the semiconductor industry

involves testing the devices at an elevated temperature, also called accelerated life testing

(ALT). This accelerated test allows the ability to create in a short period of time the same

failure modes that would take much longer to occur under normal operating conditions.

One of the most common model used to analyze device lifetime is the Arrhenius

model which uses temperature and activation energy (Ea) to predict mean time before

failure (MTBF), defined as lifetime. The following equations shows how the Arrhenius

model can be expressed to show activation energy when lifetime and temperature are

known.

72

𝑡𝑓 = 𝐴 𝑒𝑥𝑝 (𝐸𝑎

𝑘𝑇), (6.1)

where, tf is the lifetime (h); A is a scaling factor; 𝐸𝑎 is the activation energy (eV) for the

specific failure mechanism at temperature T; k is Boltzmann’s constant. Equation 6.1

could be derived as,

ln 𝑡𝑓 = 𝐸𝑎 (1

𝑘𝑇) + ln 𝐴 (6.2)

Equation 6.2 shows that there is linear relation between ln tf and 1/kT. By

measuring the lifetime at high temperatures and plotting the ln tf vs 1/kT, we can extract

the value of Ea and A, and then we are able to estimate the lifetime at any temperature.

It must be mentioned that the Arrhenius equation is applicable only if a single

mechanism is contributing to the failure of a device, because the activation energy Ea is

corresponding to one specific failure mechanism. If there is more than one failure

mechanism involved, Ea would change with temperature instead of being a constant and

the plot of ln tf vs 1/kT would display a change in slope as well. In other words, the

lifetime prediction based on Arrhenius mode is reasonable only if the plot of ln tf vs 1/kT

displays a constant slope. Although the absolute accuracy of the predicted lifetime

remains questionable, this method still provide useful information in estimating the

reliability of a device qualitatively, especially when comparing two devices under the

same accelerating conditions.

The metal-semiconductor contacts exhibit a wear-out failure mode by showing a

slow degradation in resistivity even at room temperature. We have noticed that the

contact resistivity of an alloyed Ti/Al/Ni/Au to n-GaN contact increased from 6×10-7

Ω·cm2, measured after the sample was prepared, to 6×10-6 Ω·cm2 after the sample being

73

stored at room temperature for two months. The contact degradation at high temperature

is much faster, so the lifetime is much shorter according to equation 6.1. The lifetime of

those contacts will play an important role when designing the high temperature InGaN

solar cell. In this study, the lifetime of a contact at certain temperature, also called mean

time before failure (MTBF), was defined as the time after which ρc has increased by a

factor of 10 compared with the initial value. While this can be a stringent condition for

MTBF (where the voltage drop at the leads may be negligible to disrupt the solar cell

operation) this study assumed such a criterion to evaluate the method of ALT based on

Arrhenius model.

6.2 Accelerated Life Testing of the Annealed Ni/Au to p-GaN Contact

We applied the ALT to the annealed Ni/Au to p-GaN contact, which had an initial

ρc ~ 2×10-3 Ω·cm2. Accord to the definition, the time after which the ρc increases to 2×10-

2 Ω·cm2 is the lifetime at a specified temperature. To predict the lifetime at 450 °C in air,

ALT was conducted at higher temperatures up to 525 °C in air, assuming that the

degradation mechanisms are the same in the temperature range between 450 °C and

525 °C in air.

Three samples with TLM test structures, A, B, and C, were put in the furnace

with ambient air for thermal stress at 475 °C, 500 °C, and 525 °C, respectively. The

contact samples were taken out and cooled down to room temperature every after few

hours and the contact resistivity was monitored by TLM measurements at room

temperature. Table 1 shows the values of ρc for the three samples as a function of time

under thermal stress at different temperatures. The increase of ρc as a function of time

under thermal stress is shown in Fig. 48, depicting linear increase of ρc at all the

74

temperatures. By doing a linear fit of ρc vs time under thermal stress, the lifetime of the

contact at each temperature can be extracted. The extracted lifetime at 475 °C, 500 °C,

and 525 °C in air are 69, 24, and 9 hours, respectively.

Table 1 Values of ρc for contact sample A, B, and C as a function of time under the

thermal stress at 475 °C, 500 °C, and 525 °C, respectively, in air.

A(475 °C) B(500 °C) C (525 °C)

Hours ρc

(Ω·cm2) Hours

ρc

(Ω·cm2) Hours

ρc

(Ω·cm2)

0 2.3E-03 0 1.6E-03 0 1.7E-03

4 4.0E-03 2 3.0E-03 2 5.1E-03

8 5.3E-03 4 4.2E-03 4 8.2E-03

12 6.1E-03 6 5.4E-03 6 1.1E-02

16 7.0E-03 8 6.5E-03 10 1.9E-02

Fig. 48 Increase of ρc as a function of time under thermal stress at 475 °C, 500 °C, and

525 °C, respectively, in air.

0 2 4 6 8 10 12 14 16 180.000

0.005

0.010

0.015

0.020

c

Time under thermal stress (h)

475 C

500 C

525 C

y = 0.0003x + 0.0027

y = 0.0006x + 0.0017

y = 0.0017x + 0.0016

75

With three values of temperature and the corresponding values of lifetime, ln (tf)

vs 1/kT can be plotted, shown in Fig. 49. The linear relation between ln (tf) and 1/kT

indicates that there is a single degradation mechanism in this temperature range, and thus

the Arrhenius equation is applicable. The activation energy, Ea ~2eV, and the scaling

factor, A ~7.2e-13 h, were extracted by linear fitting of the Arrhenius plot. The scaling

factor A, in equation 6.1, is different from that in the conventionally defined equation,

𝑘𝑟𝑐 = 𝐴 exp (−𝐸𝑎

𝑘𝑇), where 𝑘𝑟𝑐 is the rate coefficient of the reaction.

Fig. 49 Arrhenius plot of ln (tf) vs 1/kT in air.

With the values of Ea and A, the contact lifetime can be estimated by equation 6.1.

Fig. 50 shows the estimated lifetime versus temperature according to Arrhenius model.

The lifetime of the annealed Ni/Au to p-GaN contacts is 1.3×106 h at 300 °C, 4.4×104 h

at 350 °C, 2473 h at 400 °C, 208 h at 450 °C, respectively. The contact lifetime drops

rapidly as the operating temperature increases.

76

Fig. 50 Estimated lifetime in air versus temperature according to Arrhenius model.

In chapter 5, we claimed that the degradation of Ni/Au contact in air is probably

due to oxidation of p-GaN surface at high temperature in the oxidizing ambient. To

achieve a contrasting study and proved our hypothesis similar ALT was performed in N2

ambient. The contact degradation in N2 was accelerated by subjecting another set of

annealed Ni/Au to p-GaN contact samples, with the same initial ρc ~ 2×10-3 Ω·cm2, to

thermal stress at 500°C, 529°C, and 554°C, respectively. ρc showed much slower increase

in N2 compared with that in air. The extracted lifetime at 500°C, 529°C, and 554°C were

235, 65, and 24 hours, respectively. Ea and A were extracted as described earlier from the

plot of ln (tf) vs 1/kT. Ea = 2.33 eV while A=1.52×10-13 h. The higher activation energy

for the degradation in N2 compared with that in air indicates the contact is more stable in

N2 ambient. The predicted lifetime at 450 °C in N2 ambient is 2,628 hours, 12 times the

lifetime in air at the same temperature. The comparison of contact lifetime (hours) in air

300 400 500 60010

0

101

102

103

104

105

106

107

Estim

ate

d lifetim

e t

f (h)

Temperature (C)

tf(300 C)=1.27x10

6 h

tf(350 C)=4.36x10

4 h

tf(400 C)= 2473 h

tf(450 C)= 208 h

77

and in N2 at temperatures from 300 °C to 500 °C, is shown in Table 2, demonstrating that

the contact lifetime is much longer in N2 than that in air at high temperatures.

Fig. 51 Arrhenius plot of ln (tf) vs 1/kT in N2.

Fig. 52 Estimated lifetime in N2 versus temperature according to Arrhenius model.

300 400 500 60010

1

102

103

104

105

106

107

108

109

Estim

ate

d lifetim

e t

f (h)

Temperature (C)

tf(300 C)=4.7x10

7 h

tf(350 C)=1x10

6 h

tf(400 C)=4.2x10

4 h

tf(450 C)=2628 h

78

Table 2 Comparison of contact lifetime (hours) in air and in N2 at temperatures from

300 °C to 500 °C.

Temperature (°C) 300 350 400 450 500

In air 1.3×106 4.4×104 2,473 208 24

In N2 4.7×107 1×106 4.2×104 2,628 234

6.3 Summary

In this chapter, the accelerated life testing using Arrhenius mode has been

employed to evaluate the reliability of the ohmic contact to GaN at high temperature. We

have accomplished a quantitative analysis of lifetime for the annealed Ni/Au to p-GaN

contacts at high temperature in air and N2 ambient, respectively. At 450 °C, the lifetime

of the contact in N2 ambient is 2,628 hours, which is 12 times the lifetime in air at the

same temperature. The comparison of the contact lifetime in air and N2 indicates that the

annealed Ni/Au contact is much more stable in non-oxidizing ambient. This result is

consistent with the conclusions made in chapter 5 claiming that oxidation of p-GaN

surface is responsible for the contact degradation at high temperature in air.

79

Chapter 7

CONCLUSION AND FUTURE WORK

7.1 Conclusion

In an effort to develop ohmic contacts with low resistivity and high stability for

GaN high temperature electronics, especially the GaN-based high temperature solar cell,

we have investigated the high temperature performance and thermal stability of two

different ohmic contact schemes to n-GaN including alloyed Ti/Al/Ni/Au contacts and

non-alloyed Al/Au contacts, and two ohmic contacts to p-GaN including annealed Ni/Au

contacts and ITO contacts.

For the n-GaN contacts, temperature dependent measurements showed that ρc

didn’t change with ambient temperature while the sheet resistance of n-GaN increased

with ambient temperature. Both the alloyed Ti/Al/Ni/Au contacts and the non-alloyed

Al/Au contacts are stable for operating at temperatures up to 600 °C for at least 4 hours in

air with low contact resistivity in the order of ~10-6 Ω•cm2. The formation of TiN during

alloying is believed to be responsible for the low resistivity and high stability of the

alloyed Ti/Al/Ni/Au contacts, while the formation of Al-O-Au complex on the interface

plays an important role in reducing the contact resistivity of the Al/Au contacts and

maintaining the same at high temperatures.

The p-GaN contacts showed superior ohmic behavior at elevated temperatures.

Temperature dependent measurements showed that both ρc and sheet resistance of p-GaN

decreased by a factor of 10 as the ambient temperature increased from room temperature

to 390 °C, and expected to be reduced further at 450 °C due to the increased hole

concentration at high temperature. The annealed Ni/Au contacts, with ρc ~2×10-3 Ω cm2

80

at room temperature, showed very good stability at temperatures up to 450 °C in air. The

formation of faultless Au/p-GaN epitaxial lattice structure is believed to responsible for

the low resistivity and high stability of the Ni/Au contacts. Evidences collected during

this study suggest that the degradation of contacts at higher temperature is caused by the

oxidation of p-GaN surface. ITO contact to p-GaN, with a ρc ~1×10-2 Ω cm2 at room

temperature, is stable at 450 °C in air but it is not stable at 450 °C in N2 ambient due to

the out-diffusion of oxygen, resulting in the increase of carrier concentration in ITO and

the decrease of work function of ITO. The ITO contact, however, degrades at 500 °C in

air, most likely due to the absorption of excessive oxygen which reduces the conductivity

of ITO.

We also developed a method using ALT to quantitatively evaluate the contact

stability at high temperature. The annealed Ni/Au to p-GaN contact showed a lifetime of

2,628 hours at 450 °C in N2, which is 12 times the lifetime at 450 °C in air. The ALT

showed that the annealed Ni/Au contact was more stable for operating in N2 ambient.

In summary, we have developed stable ohmic contacts needed for the high

temperature GaN-based solar cell operating at 450 °C in both oxidizing and non-

oxidizing ambient. The temperature limit of each contact metallization scheme was

studied and possible degradation mechanism was discussed. The impressive results

discussed in this thesis demonstrate excellent performance of the ohmic contacts to GaN

at high temperature, making GaN a material of choice for high temperature applications.

7.2 Future Work

The results and discussions in this dissertation point to several interesting

directions for future work:

81

One should further analyze the interfacial layer on the Al/Au to n-GaN contacts to

understand the mechanism of reducing contact resistivity by high temperature thermal

stress. Our results in chapter 4 indicated that the formation of Al-O-Au complex was

responsible for the reduction in contact resistivity, which needs further investigation and

explanation. This finding will not only help us predict the long term stability of the Al/Au

to n-GaN contact, but also provide a way of achieving very low contact resistivity

without high temperature annealing.

A major direction to follow this work will be to test the Ni/Au to p-GaN contacts

in a pure non-oxidizing ambient, e.g. N2 and Ar. We concluded the contact degradation at

temperatures above 450 °C was due to oxidation on the interface. Effect of other non-

oxidizing ambient or baths will help understand the thermodynamics better. Studying the

humidity effect on contact degradation at high temperature is extremely important.

Humidity plays an important role on contact degradation. With thorough consideration of

temperature, ambient, and humidity effects, we will be able to estimate the lifetime of the

high temperature InGaN solar cell more accurately.

Other than studying on ALT and degradation mechanism, one should keep in

mind that the ultimate benefit of these studies won’t be accomplished until one develops

stable contacts to GaN for temperatures relevant to the high temperature applications. For

example, 600 °C is mentioned to be the sweet spot for hybrid solar towers where the

efficiency of the molten salt is at its peak around 570oC. Although we have developed

stable contacts to n-GaN for 600 °C operation, but the contacts to p-GaN discussed in this

dissertation are not stable for operation above 450 °C. Alternative metal combinations

82

and contact preparation methods need to be investigated to develop stable ohmic contact

to p-GaN for such applications.

The intrinsic GaN device (here the InGaN solar cell) has the capability to support

high temperature operations even up to 800 ºC, but one needs to verify how the cells

survive or fail at high temperatures during the operation. Therefore testing the stability of

the InGaN solar cell at high temperatures and predict the lifetime using accelerated life

testing will be a useful next step. The lifespan of a modern solar panel could be 30 years

for normal operation, but it would be much shorter at high temperature. It is very

important to know the lifespan so that we can calculate the costs and earnings of the high

temperature InGaN solar cell.

Space exploration applications rely on extreme temperature operation but for a

significantly shorter duration compared to commercial solar cells. This study is an

important addition to understand what roles GaN devices can play in space exploration

technology. The devices need not be limited to solar cells only. Power electronics, high

frequency electronics, UV detectors and sensors- all of these major fields thriving on

GaN’s superior material qualities will be benefited from high temperature contact

technologies.

83

REFERENCES

[1] P. G. Neudeck, R. S. Okojie, and L.-Y. Chen, “High-temperature electronics - a role

for wide bandgap semiconductors?,” Proc. IEEE, vol. 90, no. 6, pp. 1065–1076, Jun.

2002.

[2] H. A. Mantooth, M. M. Mojarradi, and R. W. Johnson, “Emerging capabilities in

electronics technologies for extreme environments, Part I: High temperature

electronics,” IEEE Power Electron. Soc. Newsl., vol. 18, no. 1, pp. 9–14, 2006.

[3] R. W. Johnson, J. L. Evans, P. Jacobsen, J. R. Thompson, and M. Christopher, “The

changing automotive environment: high-temperature electronics,” IEEE Trans.

Electron. Packag. Manuf., vol. 27, no. 3, pp. 164–176, Jul. 2004.

[4] W. J. Pulliam, P. M. Russler, and R. S. Fielder, “High-temperature high-bandwidth

fiber optic MEMS pressure-sensor technology for turbine engine component testing,”

in Environmental and Industrial Sensing, International Society for Optics and

Photonics, 2002. pp. 229–238, 2002.

[5] T. George, K.-A. Son, R. A. Powers, L. Y. Del Castillo, and R. Okojie, “Harsh

environment microtechnologies for NASA and terrestrial applications,” in IEEE

Sensors, 2005. pp. 1253–1258, 2005.

[6] J. C. Zolper, “A review of junction field effect transistors for high-temperature and

high-power electronics,” Solid-State Electron., vol. 42, no. 12, pp. 2153–2156, Dec.

1998.

[7] T. P. Chow, N. Ramungul, and M. Ghezzo, “Wide Bandgap Semiconductor Power

Devices,” in MRS Proceedings, 1997, vol. 483, p. 89, 1997.

[8] J. Wu, W. Walukiewicz, K. M. Yu, J. W. A. Iii, E. E. Haller, H. Lu, and W. J.

Schaff, “Small band gap bowing in In1−xGaxN alloys,” Appl. Phys. Lett., vol. 80,

no. 25, pp. 4741–4743, Jun. 2002.

[9] X. Cai, S. Zeng, and B. Zhang, “Fabrication and characterization of InGaN p-i-n

homojunction solar cell,” Appl. Phys. Lett., vol. 95, no. 17, p. 173504, Oct. 2009.

[10] X. Zheng, R.-H. Horng, D.-S. Wuu, M.-T. Chu, W.-Y. Liao, M.-H. Wu, R.-M. Lin,

and Y.-C. Lu, “High-quality InGaN∕GaN heterojunctions and their photovoltaic

effects,” Appl. Phys. Lett., vol. 93, no. 26, p. 261108, Dec. 2008.

[11] H. M. Branz, W. Regan, K. J. Gerst, J. B. Borak, and E. A. Santori, “Hybrid solar

converters for maximum exergy and inexpensive dispatchable electricity,” Energy

Environ. Sci., vol. 8, no. 11, pp. 3083–3091, Oct. 2015.

[12] A. Luque and A. Martı́, “Limiting efficiency of coupled thermal and photovoltaic

converters,” Sol. Energy Mater. Sol. Cells, vol. 58, no. 2, pp. 147–165, Jun. 1999.

84

[13] Y. Yang, W. Yang, W. Tang, and C. Sun, “High-temperature solar cell for

concentrated solar-power hybrid systems,” Appl. Phys. Lett., vol. 103, no. 8, p.

083902, Aug. 2013.

[14] O. Jani, I. Ferguson, C. Honsberg, and S. Kurtz, “Design and characterization of

GaN/InGaN solar cells,” Appl. Phys. Lett., vol. 91, no. 13, p. 132117, Sep. 2007.

[15] L. Kassamakova-Kolaklieva, R. Kakanakov, V. Cimalla, N. Hristeva, G. Lepoeva,

N. I. Kuznetsov, and K. Zekentes, “Pd-Based Ohmic Contacts to LPE 4H-SiC with

Improved Thermal Stability,” Mater. Sci. Forum, vol. 433–436, pp. 713–716, 2003.

[16] R. Kakanakov, L. Kasamakova-Kolaklieva, N. Hristeva, G. Lepoeva, J. B. Gomes, I.

Avramova, and T. Marinova, “High temperature and high power stability

investigation of Al-based ohmic contacts to p-type 4H-SiC,” in 27th International

Spring Seminar on Electronics Technology: Meeting the Challenges of Electronics

Technology Progress, 2004, vol. 3, pp. 543–546, 2004.

[17] L. Kolaklieva, R. Kakanakov, G. Lepoeva, J. B. Gomes, and T. Marinova,

“Au/Ti/Al contacts to SiC for power applications: electrical, chemical and thermal

properties,” in 24th International Conference on Microelectronics, 2004, vol. 2, pp.

421–424, 2004.

[18] R. S. Okojie, D. Lukco, Y. L. Chen, and D. J. Spry, “Reliability assessment of

Ti/TaSi2/Pt ohmic contacts on SiC after 1000 h at 600 °C,” J. Appl. Phys., vol. 91,

no. 10, pp. 6553–6559, May 2002.

[19] R. Maboudian, C. Carraro, D. G. Senesky, and C. S. Roper, “Advances in silicon

carbide science and technology at the micro- and nanoscales,” J. Vac. Sci. Technol.

A, vol. 31, no. 5, p. 050805, Sep. 2013.

[20] P. Hacke, T. Detchprohm, K. Hiramatsu, and N. Sawaki, “Schottky barrier on n‐type GaN grown by hydride vapor phase epitaxy,” Appl. Phys. Lett., vol. 63, no. 19,

pp. 2676–2678, Nov. 1993.

[21] J. D. Guo, F. M. Pan, M. S. Feng, R. J. Guo, P. F. Chou, and C. Y. Chang,

“Schottky contact and the thermal stability of Ni on n‐type GaN,” J. Appl. Phys., vol.

80, no. 3, pp. 1623–1627, Aug. 1996.

[22] S. C. Binari, H. B. Dietrich, G. Kelner, L. B. Rowland, K. Doverspike, and D. K.

Gaskill, “Electrical characterisation of Ti Schottky barriers can n-type GaN,”

Electron. Lett., vol. 30, no. 11, pp. 909–911, May 1994.

[23] A. C. Schmitz, A. T. Ping, M. A. Khan, Q. Chen, J. W. Yang, and I. Adesida,

“Schottky barrier properties of various metals on n-type GaN,” Semicond. Sci.

Technol., vol. 11, no. 10, p. 1464, 1996.

85

[24] J. D. Guo, M. S. Feng, R. J. Guo, F. M. Pan, and C. Y. Chang, “Study of Schottky

barriers on n‐type GaN grown by low‐pressure metalorganic chemical vapor

deposition,” Appl. Phys. Lett., vol. 67, no. 18, pp. 2657–2659, Oct. 1995.

[25] T. Mori, T. Kozawa, T. Ohwaki, Y. Taga, S. Nagai, S. Yamasaki, S. Asami, N.

Shibata, and M. Koike, “Schottky barriers and contact resistances on p‐type GaN,”

Appl. Phys. Lett., vol. 69, no. 23, pp. 3537–3539, Dec. 1996.

[26] Y.-F. Wu, W.-N. Jiang, B. P. Keller, S. Keller, D. Kapolnek, S. P. Denbaars, U. K.

Mishra, and B. Wilson, “Low resistance ohmic contact to n-GaN with a separate

layer method,” Solid-State Electron., vol. 41, no. 2, pp. 165–168, Feb. 1997.

[27] B. P. Luther, S. E. Mohney, T. N. Jackson, M. A. Khan, Q. Chen, and J. W. Yang,

“Investigation of the mechanism for Ohmic contact formation in Al and Ti/Al

contacts to n-type GaN,” Appl. Phys. Lett., vol. 70, no. 1, pp. 57–59, Jan. 1997.

[28] M. E. Lin, Z. Ma, F. Y. Huang, Z. F. Fan, L. H. Allen, and H. Morkoç, “Low

resistance ohmic contacts on wide band‐gap GaN,” Appl. Phys. Lett., vol. 64, no. 8,

pp. 1003–1005, Feb. 1994.

[29] A. Motayed, R. Bathe, M. C. Wood, O. S. Diouf, R. D. Vispute, and S. N.

Mohammad, “Electrical, thermal, and microstructural characteristics of Ti/Al/Ti/Au

multilayer Ohmic contacts to n-type GaN,” J. Appl. Phys., vol. 93, no. 2, pp. 1087–

1094, Jan. 2003.

[30] S. Ruvimov, Z. Liliental‐Weber, J. Washburn, K. J. Duxstad, E. E. Haller, Z.-F. Fan,

S. N. Mohammad, W. Kim, A. E. Botchkarev, and H. Morkoç, “Microstructure of

Ti/Al and Ti/Al/Ni/Au Ohmic contacts for n‐GaN,” Appl. Phys. Lett., vol. 69, no. 11,

pp. 1556–1558, Sep. 1996.

[31] B. P. Luther, J. M. DeLucca, S. E. Mohney, and R. F. K. Jr, “Analysis of a thin AlN

interfacial layer in Ti/Al and Pd/Al ohmic contacts to n-type GaN,” Appl. Phys. Lett.,

vol. 71, no. 26, pp. 3859–3861, Dec. 1997.

[32] Z. Fan, S. N. Mohammad, W. Kim, Ö. Aktas, A. E. Botchkarev, and H. Morkoç,

“Very low resistance multilayer Ohmic contact to n‐GaN,” Appl. Phys. Lett., vol. 68,

no. 12, pp. 1672–1674, Mar. 1996.

[33] N. A. Papanicolaou, M. V. Rao, J. Mittereder, and W. T. Anderson, “Reliable Ti/Al

and Ti/Al/Ni/Au ohmic contacts to n-type GaN formed by vacuum annealing,” J.

Vac. Sci. Technol. B, vol. 19, no. 1, pp. 261–267, Jan. 2001.

[34] A. N. Bright, P. J. Thomas, M. Weyland, D. M. Tricker, C. J. Humphreys, and R.

Davies, “Correlation of contact resistance with microstructure for

Au/Ni/Al/Ti/AlGaN/GaN ohmic contacts using transmission electron microscopy,”

J. Appl. Phys., vol. 89, no. 6, pp. 3143–3150, Mar. 2001.

86

[35] C.-T. Lee and H.-W. Kao, “Long-term thermal stability of Ti/Al/Pt/Au Ohmic

contacts to n-type GaN,” Appl. Phys. Lett., vol. 76, no. 17, pp. 2364–2366, Apr.

2000.

[36] Q. Z. Liu and S. S. Lau, “A review of the metal–GaN contact technology,” Solid-

State Electron., vol. 42, no. 5, pp. 677–691, May 1998.

[37] E. F. Chor, D. Zhang, H. Gong, G. L. Chen, and T. Y. F. Liew, “Electrical

characterization and metallurgical analysis of Pd-containing multilayer contacts on

GaN,” J. Appl. Phys., vol. 90, no. 3, pp. 1242–1249, Aug. 2001.

[38] V. Kumar, L. Zhou, D. Selvanathan, and I. Adesida, “Thermally-stable low-

resistance Ti/Al/Mo/Au multilayer ohmic contacts on n–GaN,” J. Appl. Phys., vol.

92, no. 3, pp. 1712–1714, Aug. 2002.

[39] S. N. Mohammad, “Contact mechanisms and design principles for alloyed ohmic

contacts to n-GaN,” J. Appl. Phys., vol. 95, no. 12, pp. 7940–7953, Jun. 2004.

[40] Z. X. Qin, Z. Z. Chen, Y. Z. Tong, X. M. Ding, X. D. Hu, T. J. Yu, and G. Y. Zhang,

“Study of Ti/Au, Ti/Al/Au, and Ti/Al/Ni/Au ohmic contacts to n-GaN,” Appl. Phys.

A, vol. 78, no. 5, pp. 729–731, Jan. 2003.

[41] M. Hou and D. G. Senesky, “Operation of ohmic Ti/Al/Pt/Au multilayer contacts to

GaN at 600 °C in air,” Appl. Phys. Lett., vol. 105, no. 8, p. 081905, Aug. 2014.

[42] F. Iucolano, F. Roccaforte, A. Alberti, C. Bongiorno, S. D. Franco, and V. Raineri,

“Temperature dependence of the specific resistance in Ti/Al/Ni/Au contacts on n-

type GaN,” J. Appl. Phys., vol. 100, no. 12, p. 123706, Dec. 2006.

[43] W. Götz, N. M. Johnson, J. Walker, D. P. Bour, and R. A. Street, “Activation of

acceptors in Mg‐doped GaN grown by metalorganic chemical vapor deposition,”

Appl. Phys. Lett., vol. 68, no. 5, pp. 667–669, Jan. 1996.

[44] J. K. Sheu, Y. K. Su, G. C. Chi, P. L. Koh, M. J. Jou, C. M. Chang, C. C. Liu, and

W. C. Hung, “High-transparency Ni/Au ohmic contact to p-type GaN,” Appl. Phys.

Lett., vol. 74, no. 16, pp. 2340–2342, Apr. 1999.

[45] J.-K. Ho, C.-S. Jong, C. C. Chiu, C.-N. Huang, C.-Y. Chen, and K.-K. Shih, “Low-

resistance ohmic contacts to p-type GaN,” Appl. Phys. Lett., vol. 74, no. 9, pp.

1275–1277, Mar. 1999.

[46] J.-K. Ho, C.-S. Jong, C. C. Chiu, C.-N. Huang, K.-K. Shih, L.-C. Chen, F.-R. Chen,

and J.-J. Kai, “Low-resistance ohmic contacts to p-type GaN achieved by the

oxidation of Ni/Au films,” J. Appl. Phys., vol. 86, no. 8, pp. 4491–4497, Oct. 1999.

87

[47] L.-C. Chen, J.-K. Ho, C.-S. Jong, C. C. Chiu, K.-K. Shih, F.-R. Chen, J.-J. Kai, and

L. Chang, “Oxidized Ni/Pt and Ni/Au ohmic contacts to p-type GaN,” Appl. Phys.

Lett., vol. 76, no. 25, pp. 3703–3705, Jun. 2000.

[48] S. Zhao, Y. Shi, H. Li, and Q. He, “The improvement of low-resistance and high-

transmission ohmic contact to p-GaN by Zn+ implantation,” Nucl. Instrum. Methods

Phys. Res. Sect. B Beam Interact. Mater. At., vol. 268, no. 9, pp. 1435–1439, May

2010.

[49] H. Omiya, F. A. Ponce, H. Marui, S. Tanaka, and T. Mukai, “Atomic arrangement

at the Au/p-GaN interface in low-resistance contacts,” Appl. Phys. Lett., vol. 85, no.

25, pp. 6143–6145, Dec. 2004.

[50] J. K. Sheu, Y. K. Su, G. C. Chi, W. C. Chen, C. Y. Chen, C. N. Huang, J. M. Hong,

Y. C. Yu, C. W. Wang, and E. K. Lin, “The effect of thermal annealing on the

Ni/Au contact of p-type GaN,” J. Appl. Phys., vol. 83, no. 6, pp. 3172–3175, Mar.

1998.

[51] C.-S. Lee, Y.-J. Lin, and C.-T. Lee, “Investigation of oxidation mechanism for

ohmic formation in Ni/Au contacts to p-type GaN layers,” Appl. Phys. Lett., vol. 79,

no. 23, pp. 3815–3817, Dec. 2001.

[52] H. W. Jang, S. Y. Kim, and J.-L. Lee, “Mechanism for Ohmic contact formation of

oxidized Ni/Au on p-type GaN,” J. Appl. Phys., vol. 94, no. 3, pp. 1748–1752, Aug.

2003.

[53] C. Y. Hu, Z. B. Ding, Z. X. Qin, Z. Z. Chen, K. Xu, Z. J. Yang, B. Shen, S. D. Yao,

and G. Y. Zhang, “Ohmic contact mechanism of Ni/Au contact to p-type GaN

studied by Rutherford backscattering spectrometry,” J. Cryst. Growth, vol. 298, pp.

808–810, Jan. 2007.

[54] J. Smalc-Koziorowska, S. Grzanka, E. Litwin-Staszewska, R. Piotrzkowski, G.

Nowak, M. Leszczynski, P. Perlin, E. Talik, J. Kozubowski, and S. Krukowski,

“Ni–Au contacts to p-type GaN – Structure and properties,” Solid-State Electron.,

vol. 54, no. 7, pp. 701–709, Jul. 2010.

[55] G. Greco, P. Prystawko, M. Leszczyński, R. L. Nigro, V. Raineri, and F. Roccaforte,

“Electro-structural evolution and Schottky barrier height in annealed Au/Ni contacts

onto p-GaN,” J. Appl. Phys., vol. 110, no. 12, p. 123703, Dec. 2011.

[56] X. J. Li, D. G. Zhao, D. S. Jiang, Z. S. Liu, P. Chen, J. J. Zhu, L. C. Le, J. Yang, X.

G. He, S. M. Zhang, B. S. Zhang, J. P. Liu, and H. Yang, “The significant effect of

the thickness of Ni film on the performance of the Ni/Au Ohmic contact to p-GaN,”

J. Appl. Phys., vol. 116, no. 16, p. 163708, Oct. 2014.

[57] Y. Koide, T. Maeda, T. Kawakami, S. Fujita, T. Uemura, N. Shibata, and M.

Murakami, “Effects of annealing in an oxygen ambient on electrical properties of

88

ohmic contacts to p-type GaN,” J. Electron. Mater., vol. 28, no. 3, pp. 341–346,

Mar. 1999.

[58] T. Maeda, Y. Koide, and M. Murakami, “Effects of NiO on electrical properties of

NiAu-based ohmic contacts for p-type GaN,” Appl. Phys. Lett., vol. 75, no. 26, pp.

4145–4147, Dec. 1999.

[59] H. Ishikawa, S. Kobayashi, Y. Koide, S. Yamasaki, S. Nagai, J. Umezaki, M. Koike,

and M. Murakami, “Effects of surface treatments and metal work functions on

electrical properties at p-GaN/metal interfaces,” J. Appl. Phys., vol. 81, no. 3, pp.

1315–1322, Feb. 1997.

[60] G. K. Reeves and H. B. Harrison, “Obtaining the specific contact resistance from

transmission line model measurements,” IEEE Electron Device Lett., vol. 3, no. 5,

pp. 111–113, May 1982.

[61] D. K. Schroder, Semiconductor Material and Device Characterization. John Wiley

& Sons, 2006.

[62] D. S. Stone, K. B. Yoder, and W. D. Sproul, “Hardness and elastic modulus of TiN

based on continuous indentation technique and new correlation,” J. Vac. Sci.

Technol. A, vol. 9, no. 4, pp. 2543–2547, Jul. 1991.

[63] H. O. Pierson, Handbook of Refractory Carbides & Nitrides: Properties,

Characteristics, Processing and Apps. Elsevier Science, 2013.

[64] S. Nakamura, M. Senoh, and T. Mukai, “High‐power InGaN/GaN double‐heterostructure violet light emitting diodes,” Appl. Phys. Lett., vol. 62, no. 19, pp.

2390–2392, May 1993.

[65] D. Steigerwald, S. Rudaz, H. Liu, R. S. Kern, W. Götz, and R. Fletcher, “III–V

Nitride semiconductors for high-performance blue and green light-emitting devices,”

JOM, vol. 49, no. 9, pp. 18–23, Sep. 1997.

[66] D. Sawdai, D. Pavlidis, and D. Cui, “Enhanced transmission line model structures

for accurate resistance evaluation of small-size contacts and for more reliable

fabrication,” IEEE Trans. Electron Devices, vol. 46, no. 7, pp. 1302–1311, Jul. 1999.

[67] I. Daumiller, C. Kirchner, M. Kamp, K. J. Ebeling, and E. Kohn, “Evaluation of the

temperature stability of AlGaN/GaN heterostructure FETs,” IEEE Electron Device

Lett., vol. 20, no. 9, pp. 448–450, Sep. 1999.

[68] L. L. Smith, R. F. Davis, M. J. Kim, R. W. Carpenter, and Y. Huang,

“Microstructure, electrical properties, and thermal stability of Au-based ohmic

contacts to p-GaN,” J. Mater. Res., vol. 12, no. 09, pp. 2249–2254, Sep. 1997.

89

[69] G. Kaune, M. A. Ruderer, E. Metwalli, W. Wang, S. Couet, K. Schlage, R.

Röhlsberger, S. V. Roth, and P. Müller-Buschbaum, “In Situ GISAXS Study of

Gold Film Growth on Conducting Polymer Films,” ACS Appl. Mater. Interfaces, vol.

1, no. 2, pp. 353–360, Feb. 2009.

[70] T. Hossain, D. Wei, N. Nepal, N. Y. Garces, J. K. Hite, H. M. Meyer, C. R. Eddy, T.

Baker, A. Mayo, J. Schmitt, and J. H. Edgar, “Insulating gallium oxide layer

produced by thermal oxidation of gallium-polar GaN,” Phys. Status Solidi C, vol. 11,

no. 3–4, pp. 565–568, Feb. 2014.

[71] H. S. Oon and K. Y. Cheong, “Recent development of gallium oxide thin film on

GaN,” Mater. Sci. Semicond. Process., vol. 16, no. 5, pp. 1217–1231, Oct. 2013.

[72] S. D. Wolter, S. E. Mohney, H. Venugopalan, A. E. Wickenden, and D. D. Koleske,

“Kinetic Study of the Oxidation of Gallium Nitride in Dry Air,” J. Electrochem.

Soc., vol. 145, no. 2, pp. 629–632, Feb. 1998.

[73] S. D. Wolter, B. P. Luther, D. L. Waltemyer, C. Önneby, S. E. Mohney, and R. J.

Molnar, “X-ray photoelectron spectroscopy and x-ray diffraction study of the

thermal oxide on gallium nitride,” Appl. Phys. Lett., vol. 70, no. 16, pp. 2156–2158,

Apr. 1997.

[74] S. Ishibashi, Y. Higuchi, Y. Ota, and K. Nakamura, “Low resistivity indium–tin

oxide transparent conductive films. II. Effect of sputtering voltage on electrical

property of films,” J. Vac. Sci. Technol. A, vol. 8, no. 3, pp. 1403–1406, May 1990.

[75] T. Ishida, H. Kobayashi, and Y. Nakato, “Structures and properties of electron‐beam‐evaporated indium tin oxide films as studied by x‐ray photoelectron

spectroscopy and work‐function measurements,” J. Appl. Phys., vol. 73, no. 9, pp.

4344–4350, May 1993.

[76] H. Kim, C. M. Gilmore, A. Piqué, J. S. Horwitz, H. Mattoussi, H. Murata, Z. H.

Kafafi, and D. B. Chrisey, “Electrical, optical, and structural properties of indium–

tin–oxide thin films for organic light-emitting devices,” J. Appl. Phys., vol. 86, no.

11, pp. 6451–6461, Dec. 1999.

[77] Y. Choi and H. Kim, “Surface Fermi level pinning and carrier transport of indium-

tin-oxide Ohmic contact to p-type GaN,” J. Alloys Compd., vol. 533, pp. 15–18,

Aug. 2012.

[78] J. K. Kim, S. Chhajed, M. F. Schubert, E. F. Schubert, A. J. Fischer, M. H.

Crawford, J. Cho, H. Kim, and C. Sone, “Light-Extraction Enhancement of GaInN

Light-Emitting Diodes by Graded-Refractive-Index Indium Tin Oxide Anti-

Reflection Contact,” Adv. Mater., vol. 20, no. 4, pp. 801–804, Feb. 2008.

90

[79] D. W. Kim, Y. J. Sung, J. W. Park, and G. Y. Yeom, “A study of transparent indium

tin oxide (ITO) contact to p-GaN,” Thin Solid Films, vol. 398–399, pp. 87–92, Nov.

2001.

[80] T. Margalith, O. Buchinsky, D. A. Cohen, A. C. Abare, M. Hansen, S. P. DenBaars,

and L. A. Coldren, “Indium tin oxide contacts to gallium nitride optoelectronic

devices,” Appl. Phys. Lett., vol. 74, no. 26, pp. 3930–3932, Jun. 1999.

[81] D. C. Paine, T. Whitson, D. Janiac, R. Beresford, C. O. Yang, and B. Lewis, “A

study of low temperature crystallization of amorphous thin film indium–tin–oxide,”

J. Appl. Phys., vol. 85, no. 12, pp. 8445–8450, Jun. 1999.

[82] V. S. Reddy, K. Das, A. Dhar, and S. K. Ray, “The effect of substrate temperature

on the properties of ITO thin films for OLED applications,” Semicond. Sci. Technol.,

vol. 21, no. 12, p. 1747, 2006.

[83] W.-L. Hsu, Y.-H. Pai, F.-S. Meng, C.-W. Liu, and G.-R. Lin, “Nanograin crystalline

transformation enhanced UV transparency of annealing refined indium tin oxide

film,” Appl. Phys. Lett., vol. 94, no. 23, p. 231906, Jun. 2009.

[84] T. Minami, “Transparent conducting oxide semiconductors for transparent

electrodes,” Semicond. Sci. Technol., vol. 20, no. 4, pp. S35–S44, Mar. 2005.

[85] T. Minami, “New n-Type Transparent Conducting Oxides,” MRS Bull., vol. 25, no.

08, pp. 38–44, Aug. 2000.

[86] T. Minami, “Transparent and conductive multicomponent oxide films prepared by

magnetron sputtering,” J. Vac. Sci. Technol. A, vol. 17, no. 4, pp. 1765–1772, Jul.

1999.

[87] M. Yamaguchi, A. Ide-Ektessabi, H. Nomura, and N. Yasui, “Characteristics of

indium tin oxide thin films prepared using electron beam evaporation,” Thin Solid

Films, vol. 447–448, pp. 115–118, Jan. 2004.

developing ohmic contacts to gallium nitride · 2016-10-12 · accelerated life testing (alt) was...

Documents